Bump version to 0.9.80

1. Add a cron mode to avoid exit(1) during cronjobs/timers
2. Revamp the remote_mount settings into auto_mount
   This removes a lot of unnecessary complexity while giving the
   administrator more flexibility in what they want to execute to mount
   a filesystem and how. The naming reflects the goal but the possible
   commands are arbitrary.

.version

@@ -1 +1 @@
-0.9.78
+0.9.80

CHANGELOG.md

@@ -1,5 +1,23 @@
 ## PVC Changelog
 
+###### [v0.9.80](https://github.com/parallelvirtualcluster/pvc/releases/tag/v0.9.80)
+
+  * [CLI] Improves CLI performance by not loading "pkg_resources" until needed
+  * [CLI] Improves the output of the audit log (full command paths)
+  * [Node Daemon/API Daemon] Moves the sample YAML configurations to /usr/share/pvc instead of /etc/pvc and cleans up the old locations automatically
+  * [CLI] Adds VM autobackup functionality to automate VM backup/retention and scheduling
+  * [CLI] Handles the internal store in a better way to ensure CLI can be used as a module properly
+
+###### [v0.9.79](https://github.com/parallelvirtualcluster/pvc/releases/tag/v0.9.79)
+
+  **API Changes**: New endpoints /vm/{vm}/backup, /vm/{vm}/restore
+
+  * [CLI Client] Fixes some storage pool help text messages
+  * [Node Daemon] Increases the IPMI monitoring plugin timeout
+  * [All] Adds support for VM backups, including creation, removal, and restore
+  * [Repository] Fixes shebangs in scripts to be consistent
+  * [Daemon Library] Improves the handling of VM list arguments (default None)
+
 ###### [v0.9.78](https://github.com/parallelvirtualcluster/pvc/releases/tag/v0.9.78)
 
   * [API, Client CLI] Fixes several bugs around image uploads; adds a new query parameter for non-raw images

api-daemon/pvcapid/Daemon.py

@@ -27,7 +27,7 @@ from ssl import SSLContext, TLSVersion
 from distutils.util import strtobool as dustrtobool
 
 # Daemon version
-version = "0.9.78"
+version = "0.9.80"
 
 # API version
 API_VERSION = 1.0

api-daemon/pvcapid/flaskapi.py

@@ -2140,7 +2140,7 @@ class API_VM_Locks(Resource):
 api.add_resource(API_VM_Locks, "/vm/<vm>/locks")
 
 
-# /vm/<vm</console
+# /vm/<vm>/console
 class API_VM_Console(Resource):
     @RequestParser([{"name": "lines"}])
     @Authenticator
@@ -2293,6 +2293,202 @@ class API_VM_Device(Resource):
 api.add_resource(API_VM_Device, "/vm/<vm>/device")
 
 
+# /vm/<vm>/backup
+class API_VM_Backup(Resource):
+    @RequestParser(
+        [
+            {
+                "name": "backup_path",
+                "required": True,
+                "helptext": "A local filesystem path on the primary coordinator must be specified",
+            },
+            {
+                "name": "incremental_parent",
+                "required": False,
+            },
+            {
+                "name": "retain_snapshot",
+                "required": False,
+            },
+        ]
+    )
+    @Authenticator
+    def post(self, vm, reqargs):
+        """
+        Create a backup of {vm} and its volumes to a local primary coordinator filesystem path
+        ---
+        tags:
+          - vm
+        parameters:
+          - in: query
+            name: backup_path
+            type: string
+            required: true
+            description: A local filesystem path on the primary coordinator to store the backup
+          - in: query
+            name: incremental_parent
+            type: string
+            required: false
+            description: A previous backup datestamp to use as an incremental parent; if unspecified a full backup is taken
+          - in: query
+            name: retain_snapshot
+            type: boolean
+            required: false
+            default: false
+            description: Whether or not to retain this backup's volume snapshots to use as a future incremental parent; full backups only
+        responses:
+          200:
+            description: OK
+            schema:
+              type: object
+              id: Message
+          400:
+            description: Execution error
+            schema:
+              type: object
+              id: Message
+          404:
+            description: Not found
+            schema:
+              type: object
+              id: Message
+        """
+        backup_path = reqargs.get("backup_path", None)
+        incremental_parent = reqargs.get("incremental_parent", None)
+        retain_snapshot = bool(strtobool(reqargs.get("retain_snapshot", "false")))
+        return api_helper.vm_backup(
+            vm, backup_path, incremental_parent, retain_snapshot
+        )
+
+    @RequestParser(
+        [
+            {
+                "name": "backup_path",
+                "required": True,
+                "helptext": "A local filesystem path on the primary coordinator must be specified",
+            },
+            {
+                "name": "backup_datestring",
+                "required": True,
+                "helptext": "A backup datestring must be specified",
+            },
+        ]
+    )
+    @Authenticator
+    def delete(self, vm, reqargs):
+        """
+        Remove a backup of {vm}, including snapshots, from a local primary coordinator filesystem path
+        ---
+        tags:
+          - vm
+        parameters:
+          - in: query
+            name: backup_path
+            type: string
+            required: true
+            description: A local filesystem path on the primary coordinator where the backup is stored
+          - in: query
+            name: backup_datestring
+            type: string
+            required: true
+            description: The backup datestring identifier (e.g. 20230102030405)
+        responses:
+          200:
+            description: OK
+            schema:
+              type: object
+              id: Message
+          400:
+            description: Execution error
+            schema:
+              type: object
+              id: Message
+          404:
+            description: Not found
+            schema:
+              type: object
+              id: Message
+        """
+        backup_path = reqargs.get("backup_path", None)
+        backup_datestring = reqargs.get("backup_datestring", None)
+        return api_helper.vm_remove_backup(vm, backup_path, backup_datestring)
+
+
+api.add_resource(API_VM_Backup, "/vm/<vm>/backup")
+
+
+# /vm/<vm>/restore
+class API_VM_Restore(Resource):
+    @RequestParser(
+        [
+            {
+                "name": "backup_path",
+                "required": True,
+                "helptext": "A local filesystem path on the primary coordinator must be specified",
+            },
+            {
+                "name": "backup_datestring",
+                "required": True,
+                "helptext": "A backup datestring must be specified",
+            },
+            {
+                "name": "retain_snapshot",
+                "required": False,
+            },
+        ]
+    )
+    @Authenticator
+    def post(self, vm, reqargs):
+        """
+        Restore a backup of {vm} and its volumes from a local primary coordinator filesystem path
+        ---
+        tags:
+          - vm
+        parameters:
+          - in: query
+            name: backup_path
+            type: string
+            required: true
+            description: A local filesystem path on the primary coordinator where the backup is stored
+          - in: query
+            name: backup_datestring
+            type: string
+            required: true
+            description: The backup datestring identifier (e.g. 20230102030405)
+          - in: query
+            name: retain_snapshot
+            type: boolean
+            required: false
+            default: true
+            description: Whether or not to retain the (parent, if incremental) volume snapshot after restore
+        responses:
+          200:
+            description: OK
+            schema:
+              type: object
+              id: Message
+          400:
+            description: Execution error
+            schema:
+              type: object
+              id: Message
+          404:
+            description: Not found
+            schema:
+              type: object
+              id: Message
+        """
+        backup_path = reqargs.get("backup_path", None)
+        backup_datestring = reqargs.get("backup_datestring", None)
+        retain_snapshot = bool(strtobool(reqargs.get("retain_snapshot", "true")))
+        return api_helper.vm_restore(
+            vm, backup_path, backup_datestring, retain_snapshot
+        )
+
+
+api.add_resource(API_VM_Restore, "/vm/<vm>/restore")
+
+
 ##########################################################
 # Client API - Network
 ##########################################################
@@ -4843,7 +5039,7 @@ class API_Storage_Ceph_Volume_Root(Resource):
             {
                 "name": "size",
                 "required": True,
-                "helptext": "A volume size in bytes (or with k/M/G/T suffix) must be specified.",
+                "helptext": "A volume size in bytes (B implied or with SI suffix k/M/G/T) must be specified.",
             },
         ]
     )
@@ -4869,7 +5065,7 @@ class API_Storage_Ceph_Volume_Root(Resource):
             name: size
             type: string
             required: true
-            description: The volume size in bytes (or with a metric suffix, i.e. k/M/G/T)
+            description: The volume size, in bytes (B implied) or with a single-character SI suffix (k/M/G/T)
         responses:
           200:
             description: OK
@@ -5122,7 +5318,7 @@ class API_Storage_Ceph_Volume_Element_Upload(Resource):
             name: file_size
             type: integer
             required: false
-            description: The size of the image file, if {image_format} is not "raw"
+            description: The size of the image file, in bytes, if {image_format} is not "raw"
         responses:
           200:
             description: OK

api-daemon/pvcapid/helper.py

@@ -470,6 +470,88 @@ def vm_define(
     return output, retcode
 
 
+@ZKConnection(config)
+def vm_backup(
+    zkhandler,
+    domain,
+    backup_path,
+    incremental_parent=None,
+    retain_snapshot=False,
+):
+    """
+    Back up a VM to a local (primary coordinator) filesystem path.
+    """
+    retflag, retdata = pvc_vm.backup_vm(
+        zkhandler,
+        domain,
+        backup_path,
+        incremental_parent,
+        retain_snapshot,
+    )
+
+    if retflag:
+        retcode = 200
+    else:
+        retcode = 400
+
+    output = {"message": retdata.replace('"', "'")}
+    return output, retcode
+
+
+@ZKConnection(config)
+def vm_remove_backup(
+    zkhandler,
+    domain,
+    source_path,
+    datestring,
+):
+    """
+    Remove a VM backup from snapshots and a local (primary coordinator) filesystem path.
+    """
+    retflag, retdata = pvc_vm.remove_backup(
+        zkhandler,
+        domain,
+        source_path,
+        datestring,
+    )
+
+    if retflag:
+        retcode = 200
+    else:
+        retcode = 400
+
+    output = {"message": retdata.replace('"', "'")}
+    return output, retcode
+
+
+@ZKConnection(config)
+def vm_restore(
+    zkhandler,
+    domain,
+    backup_path,
+    datestring,
+    retain_snapshot=False,
+):
+    """
+    Restore a VM from a local (primary coordinator) filesystem path.
+    """
+    retflag, retdata = pvc_vm.restore_vm(
+        zkhandler,
+        domain,
+        backup_path,
+        datestring,
+        retain_snapshot,
+    )
+
+    if retflag:
+        retcode = 200
+    else:
+        retcode = 400
+
+    output = {"message": retdata.replace('"', "'")}
+    return output, retcode
+
+
 @ZKConnection(config)
 def vm_attach_device(zkhandler, vm, device_spec_xml):
     """

build-and-deploy.sh

@@ -1,4 +1,4 @@
-#!/bin/bash
+#!/usr/bin/env bash
 
 # A useful script for testing out changes to PVC by building the debs and deploying them out to a
 # set of hosts automatically, including restarting the daemon (with a pause between) on the remote
@@ -36,34 +36,37 @@ echo "Preparing code (format and lint)..."
 ./lint || exit 1
 
 # Build the packages
-echo -n "Building packages... "
+echo -n "Building packages..."
 version="$( ./build-unstable-deb.sh 2>/dev/null )"
-echo "done. Package version ${version}."
+echo " done. Package version ${version}."
 
 # Install the client(s) locally
-echo -n "Installing client packages locally... "
+echo -n "Installing client packages locally..."
 $SUDO dpkg -i ../pvc-client*_${version}*.deb &>/dev/null
-echo "done".
+echo " done".
 
 for HOST in ${HOSTS[@]}; do
     echo "> Deploying packages to host ${HOST}"
-    echo -n "Copying packages... "
+    echo -n "Copying packages..."
     ssh $HOST $SUDO rm -rf /tmp/pvc &>/dev/null
     ssh $HOST mkdir /tmp/pvc &>/dev/null
     scp ../pvc-*_${version}*.deb $HOST:/tmp/pvc/ &>/dev/null
-    echo "done."
-    echo -n "Installing packages... "
+    echo " done."
+    echo -n "Installing packages..."
     ssh $HOST $SUDO dpkg -i /tmp/pvc/{pvc-client-cli,pvc-daemon-common,pvc-daemon-api,pvc-daemon-node}*.deb &>/dev/null
     ssh $HOST rm -rf /tmp/pvc &>/dev/null
-    echo "done."
-    echo -n "Restarting PVC daemons... "
+    echo " done."
+    echo -n "Restarting PVC daemons..."
     ssh $HOST $SUDO systemctl restart pvcapid &>/dev/null
     ssh $HOST $SUDO systemctl restart pvcapid-worker &>/dev/null
     ssh $HOST $SUDO systemctl restart pvcnoded &>/dev/null
-    echo "done."
-    echo -n "Waiting 30s for host to stabilize... "
-    sleep 30
-    echo "done."
+    echo " done."
+    echo -n "Waiting for node daemon to be running..."
+    while [[ $( ssh $HOST "pvc -q node list -f json ${HOST%%.*} | jq -r '.[].daemon_state'" ) != "run" ]]; do
+        sleep 5
+        echo -n "."
+    done
+    echo " done."
 done
 if [[ -z ${KEEP_ARTIFACTS} ]]; then
     rm ../pvc*_${version}*

build-stable-deb.sh

@@ -1,4 +1,4 @@
-#!/bin/sh
+#!/usr/bin/env bash
 pushd $( git rev-parse --show-toplevel ) &>/dev/null
 ver="$( head -1 debian/changelog | awk -F'[()-]' '{ print $2 }' )"
 git pull

build-unstable-deb.sh

@@ -1,4 +1,4 @@
-#!/bin/sh
+#!/usr/bin/env bash
 set -o xtrace
 exec 3>&1
 exec 1>&2

client-cli-old/pvc.py

@@ -1,33 +0,0 @@
-#!/usr/bin/env python3
-
-# pvc.py - PVC client command-line interface (stub testing interface)
-# Part of the Parallel Virtual Cluster (PVC) system
-#
-#    Copyright (C) 2018-2022 Joshua M. Boniface <joshua@boniface.me>
-#
-#    This program is free software: you can redistribute it and/or modify
-#    it under the terms of the GNU General Public License as published by
-#    the Free Software Foundation, version 3.
-#
-#    This program is distributed in the hope that it will be useful,
-#    but WITHOUT ANY WARRANTY; without even the implied warranty of
-#    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
-#    GNU General Public License for more details.
-#
-#    You should have received a copy of the GNU General Public License
-#    along with this program.  If not, see <https://www.gnu.org/licenses/>.
-#
-###############################################################################
-
-import pvc.pvc
-
-
-#
-# Main entry point
-#
-def main():
-    return pvc.pvc.cli(obj={})
-
-
-if __name__ == "__main__":
-    main()

client-cli-old/pvc/lib/ansiprint.py

@@ -1,97 +0,0 @@
-#!/usr/bin/env python3
-
-# ansiprint.py - Printing function for formatted messages
-# Part of the Parallel Virtual Cluster (PVC) system
-#
-#    Copyright (C) 2018-2022 Joshua M. Boniface <joshua@boniface.me>
-#
-#    This program is free software: you can redistribute it and/or modify
-#    it under the terms of the GNU General Public License as published by
-#    the Free Software Foundation, version 3.
-#
-#    This program is distributed in the hope that it will be useful,
-#    but WITHOUT ANY WARRANTY; without even the implied warranty of
-#    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
-#    GNU General Public License for more details.
-#
-#    You should have received a copy of the GNU General Public License
-#    along with this program.  If not, see <https://www.gnu.org/licenses/>.
-#
-###############################################################################
-
-import datetime
-
-
-# ANSII colours for output
-def red():
-    return "\033[91m"
-
-
-def blue():
-    return "\033[94m"
-
-
-def cyan():
-    return "\033[96m"
-
-
-def green():
-    return "\033[92m"
-
-
-def yellow():
-    return "\033[93m"
-
-
-def purple():
-    return "\033[95m"
-
-
-def bold():
-    return "\033[1m"
-
-
-def end():
-    return "\033[0m"
-
-
-# Print function
-def echo(message, prefix, state):
-    # Get the date
-    date = "{} - ".format(datetime.datetime.now().strftime("%Y/%m/%d %H:%M:%S.%f"))
-    endc = end()
-
-    # Continuation
-    if state == "c":
-        date = ""
-        colour = ""
-        prompt = "    "
-    # OK
-    elif state == "o":
-        colour = green()
-        prompt = ">>> "
-    # Error
-    elif state == "e":
-        colour = red()
-        prompt = ">>> "
-    # Warning
-    elif state == "w":
-        colour = yellow()
-        prompt = ">>> "
-    # Tick
-    elif state == "t":
-        colour = purple()
-        prompt = ">>> "
-    # Information
-    elif state == "i":
-        colour = blue()
-        prompt = ">>> "
-    else:
-        colour = bold()
-        prompt = ">>> "
-
-    # Append space to prefix
-    if prefix != "":
-        prefix = prefix + " "
-
-    print(colour + prompt + endc + date + prefix + message)

client-cli-old/pvc/lib/cluster.py

@@ -1,313 +0,0 @@
-#!/usr/bin/env python3
-
-# cluster.py - PVC CLI client function library, cluster management
-# Part of the Parallel Virtual Cluster (PVC) system
-#
-#    Copyright (C) 2018-2022 Joshua M. Boniface <joshua@boniface.me>
-#
-#    This program is free software: you can redistribute it and/or modify
-#    it under the terms of the GNU General Public License as published by
-#    the Free Software Foundation, version 3.
-#
-#    This program is distributed in the hope that it will be useful,
-#    but WITHOUT ANY WARRANTY; without even the implied warranty of
-#    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
-#    GNU General Public License for more details.
-#
-#    You should have received a copy of the GNU General Public License
-#    along with this program.  If not, see <https://www.gnu.org/licenses/>.
-#
-###############################################################################
-
-import json
-
-import pvc.lib.ansiprint as ansiprint
-from pvc.lib.common import call_api
-
-
-def initialize(config, overwrite=False):
-    """
-    Initialize the PVC cluster
-
-    API endpoint: GET /api/v1/initialize
-    API arguments: overwrite, yes-i-really-mean-it
-    API schema: {json_data_object}
-    """
-    params = {"yes-i-really-mean-it": "yes", "overwrite": overwrite}
-    response = call_api(config, "post", "/initialize", params=params)
-
-    if response.status_code == 200:
-        retstatus = True
-    else:
-        retstatus = False
-
-    return retstatus, response.json().get("message", "")
-
-
-def backup(config):
-    """
-    Get a JSON backup of the cluster
-
-    API endpoint: GET /api/v1/backup
-    API arguments:
-    API schema: {json_data_object}
-    """
-    response = call_api(config, "get", "/backup")
-
-    if response.status_code == 200:
-        return True, response.json()
-    else:
-        return False, response.json().get("message", "")
-
-
-def restore(config, cluster_data):
-    """
-    Restore a JSON backup to the cluster
-
-    API endpoint: POST /api/v1/restore
-    API arguments: yes-i-really-mean-it
-    API schema: {json_data_object}
-    """
-    cluster_data_json = json.dumps(cluster_data)
-
-    params = {"yes-i-really-mean-it": "yes"}
-    data = {"cluster_data": cluster_data_json}
-    response = call_api(config, "post", "/restore", params=params, data=data)
-
-    if response.status_code == 200:
-        retstatus = True
-    else:
-        retstatus = False
-
-    return retstatus, response.json().get("message", "")
-
-
-def maintenance_mode(config, state):
-    """
-    Enable or disable PVC cluster maintenance mode
-
-    API endpoint: POST /api/v1/status
-    API arguments: {state}={state}
-    API schema: {json_data_object}
-    """
-    params = {"state": state}
-    response = call_api(config, "post", "/status", params=params)
-
-    if response.status_code == 200:
-        retstatus = True
-    else:
-        retstatus = False
-
-    return retstatus, response.json().get("message", "")
-
-
-def get_info(config):
-    """
-    Get status of the PVC cluster
-
-    API endpoint: GET /api/v1/status
-    API arguments:
-    API schema: {json_data_object}
-    """
-    response = call_api(config, "get", "/status")
-
-    if response.status_code == 200:
-        return True, response.json()
-    else:
-        return False, response.json().get("message", "")
-
-
-def format_info(cluster_information, oformat):
-    if oformat == "json":
-        return json.dumps(cluster_information)
-
-    if oformat == "json-pretty":
-        return json.dumps(cluster_information, indent=4)
-
-    # Plain formatting, i.e. human-readable
-    if (
-        cluster_information.get("maintenance") == "true"
-        or cluster_information.get("cluster_health", {}).get("health", "N/A") == "N/A"
-    ):
-        health_colour = ansiprint.blue()
-    elif cluster_information.get("cluster_health", {}).get("health", 100) > 90:
-        health_colour = ansiprint.green()
-    elif cluster_information.get("cluster_health", {}).get("health", 100) > 50:
-        health_colour = ansiprint.yellow()
-    else:
-        health_colour = ansiprint.red()
-
-    ainformation = []
-
-    ainformation.append(
-        "{}PVC cluster status:{}".format(ansiprint.bold(), ansiprint.end())
-    )
-    ainformation.append("")
-
-    health_text = (
-        f"{cluster_information.get('cluster_health', {}).get('health', 'N/A')}"
-    )
-    if health_text != "N/A":
-        health_text += "%"
-    if cluster_information.get("maintenance") == "true":
-        health_text += " (maintenance on)"
-
-    ainformation.append(
-        "{}Cluster health:{}  {}{}{}".format(
-            ansiprint.purple(),
-            ansiprint.end(),
-            health_colour,
-            health_text,
-            ansiprint.end(),
-        )
-    )
-    if cluster_information.get("cluster_health", {}).get("messages"):
-        health_messages = "\n                 > ".join(
-            sorted(cluster_information["cluster_health"]["messages"])
-        )
-        ainformation.append(
-            "{}Health messages:{} > {}".format(
-                ansiprint.purple(),
-                ansiprint.end(),
-                health_messages,
-            )
-        )
-    else:
-        ainformation.append(
-            "{}Health messages:{} N/A".format(
-                ansiprint.purple(),
-                ansiprint.end(),
-            )
-        )
-
-    if oformat == "short":
-        return "\n".join(ainformation)
-
-    ainformation.append("")
-    ainformation.append(
-        "{}Primary node:{}        {}".format(
-            ansiprint.purple(), ansiprint.end(), cluster_information["primary_node"]
-        )
-    )
-    ainformation.append(
-        "{}PVC version:{}         {}".format(
-            ansiprint.purple(),
-            ansiprint.end(),
-            cluster_information.get("pvc_version", "N/A"),
-        )
-    )
-    ainformation.append(
-        "{}Cluster upstream IP:{} {}".format(
-            ansiprint.purple(), ansiprint.end(), cluster_information["upstream_ip"]
-        )
-    )
-    ainformation.append("")
-    ainformation.append(
-        "{}Total nodes:{}     {}".format(
-            ansiprint.purple(), ansiprint.end(), cluster_information["nodes"]["total"]
-        )
-    )
-    ainformation.append(
-        "{}Total VMs:{}       {}".format(
-            ansiprint.purple(), ansiprint.end(), cluster_information["vms"]["total"]
-        )
-    )
-    ainformation.append(
-        "{}Total networks:{}  {}".format(
-            ansiprint.purple(), ansiprint.end(), cluster_information["networks"]
-        )
-    )
-    ainformation.append(
-        "{}Total OSDs:{}      {}".format(
-            ansiprint.purple(), ansiprint.end(), cluster_information["osds"]["total"]
-        )
-    )
-    ainformation.append(
-        "{}Total pools:{}     {}".format(
-            ansiprint.purple(), ansiprint.end(), cluster_information["pools"]
-        )
-    )
-    ainformation.append(
-        "{}Total volumes:{}   {}".format(
-            ansiprint.purple(), ansiprint.end(), cluster_information["volumes"]
-        )
-    )
-    ainformation.append(
-        "{}Total snapshots:{} {}".format(
-            ansiprint.purple(), ansiprint.end(), cluster_information["snapshots"]
-        )
-    )
-
-    nodes_string = "{}Nodes:{} {}/{} {}ready,run{}".format(
-        ansiprint.purple(),
-        ansiprint.end(),
-        cluster_information["nodes"].get("run,ready", 0),
-        cluster_information["nodes"].get("total", 0),
-        ansiprint.green(),
-        ansiprint.end(),
-    )
-    for state, count in cluster_information["nodes"].items():
-        if state == "total" or state == "run,ready":
-            continue
-
-        nodes_string += " {}/{} {}{}{}".format(
-            count,
-            cluster_information["nodes"]["total"],
-            ansiprint.yellow(),
-            state,
-            ansiprint.end(),
-        )
-
-    ainformation.append("")
-    ainformation.append(nodes_string)
-
-    vms_string = "{}VMs:{} {}/{} {}start{}".format(
-        ansiprint.purple(),
-        ansiprint.end(),
-        cluster_information["vms"].get("start", 0),
-        cluster_information["vms"].get("total", 0),
-        ansiprint.green(),
-        ansiprint.end(),
-    )
-    for state, count in cluster_information["vms"].items():
-        if state == "total" or state == "start":
-            continue
-
-        if state in ["disable", "migrate", "unmigrate", "provision"]:
-            colour = ansiprint.blue()
-        else:
-            colour = ansiprint.yellow()
-
-        vms_string += " {}/{} {}{}{}".format(
-            count, cluster_information["vms"]["total"], colour, state, ansiprint.end()
-        )
-
-    ainformation.append("")
-    ainformation.append(vms_string)
-
-    if cluster_information["osds"]["total"] > 0:
-        osds_string = "{}Ceph OSDs:{} {}/{} {}up,in{}".format(
-            ansiprint.purple(),
-            ansiprint.end(),
-            cluster_information["osds"].get("up,in", 0),
-            cluster_information["osds"].get("total", 0),
-            ansiprint.green(),
-            ansiprint.end(),
-        )
-        for state, count in cluster_information["osds"].items():
-            if state == "total" or state == "up,in":
-                continue
-
-            osds_string += " {}/{} {}{}{}".format(
-                count,
-                cluster_information["osds"]["total"],
-                ansiprint.yellow(),
-                state,
-                ansiprint.end(),
-            )
-
-        ainformation.append("")
-        ainformation.append(osds_string)
-
-    ainformation.append("")
-    return "\n".join(ainformation)

client-cli-old/pvc/lib/common.py

@@ -1,201 +0,0 @@
-#!/usr/bin/env python3
-
-# common.py - PVC CLI client function library, Common functions
-# Part of the Parallel Virtual Cluster (PVC) system
-#
-#    Copyright (C) 2018-2022 Joshua M. Boniface <joshua@boniface.me>
-#
-#    This program is free software: you can redistribute it and/or modify
-#    it under the terms of the GNU General Public License as published by
-#    the Free Software Foundation, version 3.
-#
-#    This program is distributed in the hope that it will be useful,
-#    but WITHOUT ANY WARRANTY; without even the implied warranty of
-#    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
-#    GNU General Public License for more details.
-#
-#    You should have received a copy of the GNU General Public License
-#    along with this program.  If not, see <https://www.gnu.org/licenses/>.
-#
-###############################################################################
-
-import os
-import math
-import time
-import requests
-import click
-from urllib3 import disable_warnings
-
-
-def format_bytes(size_bytes):
-    byte_unit_matrix = {
-        "B": 1,
-        "K": 1024,
-        "M": 1024 * 1024,
-        "G": 1024 * 1024 * 1024,
-        "T": 1024 * 1024 * 1024 * 1024,
-        "P": 1024 * 1024 * 1024 * 1024 * 1024,
-    }
-    human_bytes = "0B"
-    for unit in sorted(byte_unit_matrix, key=byte_unit_matrix.get):
-        formatted_bytes = int(math.ceil(size_bytes / byte_unit_matrix[unit]))
-        if formatted_bytes < 10000:
-            human_bytes = "{}{}".format(formatted_bytes, unit)
-            break
-    return human_bytes
-
-
-def format_metric(integer):
-    integer_unit_matrix = {
-        "": 1,
-        "K": 1000,
-        "M": 1000 * 1000,
-        "B": 1000 * 1000 * 1000,
-        "T": 1000 * 1000 * 1000 * 1000,
-        "Q": 1000 * 1000 * 1000 * 1000 * 1000,
-    }
-    human_integer = "0"
-    for unit in sorted(integer_unit_matrix, key=integer_unit_matrix.get):
-        formatted_integer = int(math.ceil(integer / integer_unit_matrix[unit]))
-        if formatted_integer < 10000:
-            human_integer = "{}{}".format(formatted_integer, unit)
-            break
-    return human_integer
-
-
-class UploadProgressBar(object):
-    def __init__(self, filename, end_message="", end_nl=True):
-        file_size = os.path.getsize(filename)
-        file_size_human = format_bytes(file_size)
-        click.echo("Uploading file (total size {})...".format(file_size_human))
-
-        self.length = file_size
-        self.time_last = int(round(time.time() * 1000)) - 1000
-        self.bytes_last = 0
-        self.bytes_diff = 0
-        self.is_end = False
-
-        self.end_message = end_message
-        self.end_nl = end_nl
-        if not self.end_nl:
-            self.end_suffix = " "
-        else:
-            self.end_suffix = ""
-
-        self.bar = click.progressbar(length=self.length, show_eta=True)
-
-    def update(self, monitor):
-        bytes_cur = monitor.bytes_read
-        self.bytes_diff += bytes_cur - self.bytes_last
-        if self.bytes_last == bytes_cur:
-            self.is_end = True
-        self.bytes_last = bytes_cur
-
-        time_cur = int(round(time.time() * 1000))
-        if (time_cur - 1000) > self.time_last:
-            self.time_last = time_cur
-            self.bar.update(self.bytes_diff)
-            self.bytes_diff = 0
-
-        if self.is_end:
-            self.bar.update(self.bytes_diff)
-            self.bytes_diff = 0
-            click.echo()
-            click.echo()
-            if self.end_message:
-                click.echo(self.end_message + self.end_suffix, nl=self.end_nl)
-
-
-class ErrorResponse(requests.Response):
-    def __init__(self, json_data, status_code):
-        self.json_data = json_data
-        self.status_code = status_code
-
-    def json(self):
-        return self.json_data
-
-
-def call_api(
-    config,
-    operation,
-    request_uri,
-    headers={},
-    params=None,
-    data=None,
-    files=None,
-):
-    # Set the connect timeout to 2 seconds but extremely long (48 hour) data timeout
-    timeout = (2.05, 172800)
-
-    # Craft the URI
-    uri = "{}://{}{}{}".format(
-        config["api_scheme"], config["api_host"], config["api_prefix"], request_uri
-    )
-
-    # Craft the authentication header if required
-    if config["api_key"]:
-        headers["X-Api-Key"] = config["api_key"]
-
-    # Determine the request type and hit the API
-    disable_warnings()
-    try:
-        if operation == "get":
-            response = requests.get(
-                uri,
-                timeout=timeout,
-                headers=headers,
-                params=params,
-                data=data,
-                verify=config["verify_ssl"],
-            )
-        if operation == "post":
-            response = requests.post(
-                uri,
-                timeout=timeout,
-                headers=headers,
-                params=params,
-                data=data,
-                files=files,
-                verify=config["verify_ssl"],
-            )
-        if operation == "put":
-            response = requests.put(
-                uri,
-                timeout=timeout,
-                headers=headers,
-                params=params,
-                data=data,
-                files=files,
-                verify=config["verify_ssl"],
-            )
-        if operation == "patch":
-            response = requests.patch(
-                uri,
-                timeout=timeout,
-                headers=headers,
-                params=params,
-                data=data,
-                verify=config["verify_ssl"],
-            )
-        if operation == "delete":
-            response = requests.delete(
-                uri,
-                timeout=timeout,
-                headers=headers,
-                params=params,
-                data=data,
-                verify=config["verify_ssl"],
-            )
-    except Exception as e:
-        message = "Failed to connect to the API: {}".format(e)
-        response = ErrorResponse({"message": message}, 500)
-
-    # Display debug output
-    if config["debug"]:
-        click.echo("API endpoint: {}".format(uri), err=True)
-        click.echo("Response code: {}".format(response.status_code), err=True)
-        click.echo("Response headers: {}".format(response.headers), err=True)
-        click.echo(err=True)
-
-    # Return the response object
-    return response

client-cli-old/pvc/lib/node.py

@@ -1,709 +0,0 @@
-#!/usr/bin/env python3
-
-# node.py - PVC CLI client function library, node management
-# Part of the Parallel Virtual Cluster (PVC) system
-#
-#    Copyright (C) 2018-2022 Joshua M. Boniface <joshua@boniface.me>
-#
-#    This program is free software: you can redistribute it and/or modify
-#    it under the terms of the GNU General Public License as published by
-#    the Free Software Foundation, version 3.
-#
-#    This program is distributed in the hope that it will be useful,
-#    but WITHOUT ANY WARRANTY; without even the implied warranty of
-#    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
-#    GNU General Public License for more details.
-#
-#    You should have received a copy of the GNU General Public License
-#    along with this program.  If not, see <https://www.gnu.org/licenses/>.
-#
-###############################################################################
-
-import time
-
-import pvc.lib.ansiprint as ansiprint
-from pvc.lib.common import call_api
-
-
-#
-# Primary functions
-#
-def node_coordinator_state(config, node, action):
-    """
-    Set node coordinator state state (primary/secondary)
-
-    API endpoint: POST /api/v1/node/{node}/coordinator-state
-    API arguments: action={action}
-    API schema: {"message": "{data}"}
-    """
-    params = {"state": action}
-    response = call_api(
-        config,
-        "post",
-        "/node/{node}/coordinator-state".format(node=node),
-        params=params,
-    )
-
-    if response.status_code == 200:
-        retstatus = True
-    else:
-        retstatus = False
-
-    return retstatus, response.json().get("message", "")
-
-
-def node_domain_state(config, node, action, wait):
-    """
-    Set node domain state state (flush/ready)
-
-    API endpoint: POST /api/v1/node/{node}/domain-state
-    API arguments: action={action}, wait={wait}
-    API schema: {"message": "{data}"}
-    """
-    params = {"state": action, "wait": str(wait).lower()}
-    response = call_api(
-        config, "post", "/node/{node}/domain-state".format(node=node), params=params
-    )
-
-    if response.status_code == 200:
-        retstatus = True
-    else:
-        retstatus = False
-
-    return retstatus, response.json().get("message", "")
-
-
-def view_node_log(config, node, lines=100):
-    """
-    Return node log lines from the API (and display them in a pager in the main CLI)
-
-    API endpoint: GET /node/{node}/log
-    API arguments: lines={lines}
-    API schema: {"name":"{node}","data":"{node_log}"}
-    """
-    params = {"lines": lines}
-    response = call_api(
-        config, "get", "/node/{node}/log".format(node=node), params=params
-    )
-
-    if response.status_code != 200:
-        return False, response.json().get("message", "")
-
-    node_log = response.json()["data"]
-
-    # Shrink the log buffer to length lines
-    shrunk_log = node_log.split("\n")[-lines:]
-    loglines = "\n".join(shrunk_log)
-
-    return True, loglines
-
-
-def follow_node_log(config, node, lines=10):
-    """
-    Return and follow node log lines from the API
-
-    API endpoint: GET /node/{node}/log
-    API arguments: lines={lines}
-    API schema: {"name":"{nodename}","data":"{node_log}"}
-    """
-    # We always grab 200 to match the follow call, but only _show_ `lines` number
-    params = {"lines": 200}
-    response = call_api(
-        config, "get", "/node/{node}/log".format(node=node), params=params
-    )
-
-    if response.status_code != 200:
-        return False, response.json().get("message", "")
-
-    # Shrink the log buffer to length lines
-    node_log = response.json()["data"]
-    shrunk_log = node_log.split("\n")[-int(lines) :]
-    loglines = "\n".join(shrunk_log)
-
-    # Print the initial data and begin following
-    print(loglines, end="")
-    print("\n", end="")
-
-    while True:
-        # Grab the next line set (200 is a reasonable number of lines per half-second; any more are skipped)
-        try:
-            params = {"lines": 200}
-            response = call_api(
-                config, "get", "/node/{node}/log".format(node=node), params=params
-            )
-            new_node_log = response.json()["data"]
-        except Exception:
-            break
-        # Split the new and old log strings into constitutent lines
-        old_node_loglines = node_log.split("\n")
-        new_node_loglines = new_node_log.split("\n")
-
-        # Set the node log to the new log value for the next iteration
-        node_log = new_node_log
-
-        # Get the difference between the two sets of lines
-        old_node_loglines_set = set(old_node_loglines)
-        diff_node_loglines = [
-            x for x in new_node_loglines if x not in old_node_loglines_set
-        ]
-
-        # If there's a difference, print it out
-        if len(diff_node_loglines) > 0:
-            print("\n".join(diff_node_loglines), end="")
-            print("\n", end="")
-
-        # Wait half a second
-        time.sleep(0.5)
-
-    return True, ""
-
-
-def node_info(config, node):
-    """
-    Get information about node
-
-    API endpoint: GET /api/v1/node/{node}
-    API arguments:
-    API schema: {json_data_object}
-    """
-    response = call_api(config, "get", "/node/{node}".format(node=node))
-
-    if response.status_code == 200:
-        if isinstance(response.json(), list) and len(response.json()) != 1:
-            # No exact match, return not found
-            return False, "Node not found."
-        else:
-            # Return a single instance if the response is a list
-            if isinstance(response.json(), list):
-                return True, response.json()[0]
-            # This shouldn't happen, but is here just in case
-            else:
-                return True, response.json()
-    else:
-        return False, response.json().get("message", "")
-
-
-def node_list(
-    config, limit, target_daemon_state, target_coordinator_state, target_domain_state
-):
-    """
-    Get list information about nodes (limited by {limit})
-
-    API endpoint: GET /api/v1/node
-    API arguments: limit={limit}
-    API schema: [{json_data_object},{json_data_object},etc.]
-    """
-    params = dict()
-    if limit:
-        params["limit"] = limit
-    if target_daemon_state:
-        params["daemon_state"] = target_daemon_state
-    if target_coordinator_state:
-        params["coordinator_state"] = target_coordinator_state
-    if target_domain_state:
-        params["domain_state"] = target_domain_state
-
-    response = call_api(config, "get", "/node", params=params)
-
-    if response.status_code == 200:
-        return True, response.json()
-    else:
-        return False, response.json().get("message", "")
-
-
-#
-# Output display functions
-#
-def getOutputColours(node_information):
-    node_health = node_information.get("health", "N/A")
-    if isinstance(node_health, int):
-        if node_health <= 50:
-            health_colour = ansiprint.red()
-        elif node_health <= 90:
-            health_colour = ansiprint.yellow()
-        elif node_health <= 100:
-            health_colour = ansiprint.green()
-        else:
-            health_colour = ansiprint.blue()
-    else:
-        health_colour = ansiprint.blue()
-
-    if node_information["daemon_state"] == "run":
-        daemon_state_colour = ansiprint.green()
-    elif node_information["daemon_state"] == "stop":
-        daemon_state_colour = ansiprint.red()
-    elif node_information["daemon_state"] == "shutdown":
-        daemon_state_colour = ansiprint.yellow()
-    elif node_information["daemon_state"] == "init":
-        daemon_state_colour = ansiprint.yellow()
-    elif node_information["daemon_state"] == "dead":
-        daemon_state_colour = ansiprint.red() + ansiprint.bold()
-    else:
-        daemon_state_colour = ansiprint.blue()
-
-    if node_information["coordinator_state"] == "primary":
-        coordinator_state_colour = ansiprint.green()
-    elif node_information["coordinator_state"] == "secondary":
-        coordinator_state_colour = ansiprint.blue()
-    else:
-        coordinator_state_colour = ansiprint.cyan()
-
-    if node_information["domain_state"] == "ready":
-        domain_state_colour = ansiprint.green()
-    else:
-        domain_state_colour = ansiprint.blue()
-
-    if node_information["memory"]["allocated"] > node_information["memory"]["total"]:
-        mem_allocated_colour = ansiprint.yellow()
-    else:
-        mem_allocated_colour = ""
-
-    if node_information["memory"]["provisioned"] > node_information["memory"]["total"]:
-        mem_provisioned_colour = ansiprint.yellow()
-    else:
-        mem_provisioned_colour = ""
-
-    return (
-        health_colour,
-        daemon_state_colour,
-        coordinator_state_colour,
-        domain_state_colour,
-        mem_allocated_colour,
-        mem_provisioned_colour,
-    )
-
-
-def format_info(node_information, long_output):
-    (
-        health_colour,
-        daemon_state_colour,
-        coordinator_state_colour,
-        domain_state_colour,
-        mem_allocated_colour,
-        mem_provisioned_colour,
-    ) = getOutputColours(node_information)
-
-    # Format a nice output; do this line-by-line then concat the elements at the end
-    ainformation = []
-    # Basic information
-    ainformation.append(
-        "{}Name:{}                  {}".format(
-            ansiprint.purple(),
-            ansiprint.end(),
-            node_information["name"],
-        )
-    )
-    ainformation.append(
-        "{}PVC Version:{}           {}".format(
-            ansiprint.purple(),
-            ansiprint.end(),
-            node_information["pvc_version"],
-        )
-    )
-
-    node_health = node_information.get("health", "N/A")
-    if isinstance(node_health, int):
-        node_health_text = f"{node_health}%"
-    else:
-        node_health_text = node_health
-    ainformation.append(
-        "{}Health:{}                {}{}{}".format(
-            ansiprint.purple(),
-            ansiprint.end(),
-            health_colour,
-            node_health_text,
-            ansiprint.end(),
-        )
-    )
-
-    node_health_details = node_information.get("health_details", [])
-    if long_output:
-        node_health_messages = "\n                       ".join(
-            [f"{plugin['name']}: {plugin['message']}" for plugin in node_health_details]
-        )
-    else:
-        node_health_messages = "\n                       ".join(
-            [
-                f"{plugin['name']}: {plugin['message']}"
-                for plugin in node_health_details
-                if int(plugin.get("health_delta", 0)) > 0
-            ]
-        )
-
-    if len(node_health_messages) > 0:
-        ainformation.append(
-            "{}Health Plugin Details:{} {}".format(
-                ansiprint.purple(), ansiprint.end(), node_health_messages
-            )
-        )
-    ainformation.append("")
-
-    ainformation.append(
-        "{}Daemon State:{}          {}{}{}".format(
-            ansiprint.purple(),
-            ansiprint.end(),
-            daemon_state_colour,
-            node_information["daemon_state"],
-            ansiprint.end(),
-        )
-    )
-    ainformation.append(
-        "{}Coordinator State:{}     {}{}{}".format(
-            ansiprint.purple(),
-            ansiprint.end(),
-            coordinator_state_colour,
-            node_information["coordinator_state"],
-            ansiprint.end(),
-        )
-    )
-    ainformation.append(
-        "{}Domain State:{}          {}{}{}".format(
-            ansiprint.purple(),
-            ansiprint.end(),
-            domain_state_colour,
-            node_information["domain_state"],
-            ansiprint.end(),
-        )
-    )
-    if long_output:
-        ainformation.append("")
-        ainformation.append(
-            "{}Architecture:{}          {}".format(
-                ansiprint.purple(), ansiprint.end(), node_information["arch"]
-            )
-        )
-        ainformation.append(
-            "{}Operating System:{}      {}".format(
-                ansiprint.purple(), ansiprint.end(), node_information["os"]
-            )
-        )
-        ainformation.append(
-            "{}Kernel Version:{}        {}".format(
-                ansiprint.purple(), ansiprint.end(), node_information["kernel"]
-            )
-        )
-    ainformation.append("")
-    ainformation.append(
-        "{}Active VM Count:{}       {}".format(
-            ansiprint.purple(), ansiprint.end(), node_information["domains_count"]
-        )
-    )
-    ainformation.append(
-        "{}Host CPUs:{}             {}".format(
-            ansiprint.purple(), ansiprint.end(), node_information["vcpu"]["total"]
-        )
-    )
-    ainformation.append(
-        "{}vCPUs:{}                 {}".format(
-            ansiprint.purple(), ansiprint.end(), node_information["vcpu"]["allocated"]
-        )
-    )
-    ainformation.append(
-        "{}Load:{}                  {}".format(
-            ansiprint.purple(), ansiprint.end(), node_information["load"]
-        )
-    )
-    ainformation.append(
-        "{}Total RAM (MiB):{}       {}".format(
-            ansiprint.purple(), ansiprint.end(), node_information["memory"]["total"]
-        )
-    )
-    ainformation.append(
-        "{}Used RAM (MiB):{}        {}".format(
-            ansiprint.purple(), ansiprint.end(), node_information["memory"]["used"]
-        )
-    )
-    ainformation.append(
-        "{}Free RAM (MiB):{}        {}".format(
-            ansiprint.purple(), ansiprint.end(), node_information["memory"]["free"]
-        )
-    )
-    ainformation.append(
-        "{}Allocated RAM (MiB):{}   {}{}{}".format(
-            ansiprint.purple(),
-            ansiprint.end(),
-            mem_allocated_colour,
-            node_information["memory"]["allocated"],
-            ansiprint.end(),
-        )
-    )
-    ainformation.append(
-        "{}Provisioned RAM (MiB):{} {}{}{}".format(
-            ansiprint.purple(),
-            ansiprint.end(),
-            mem_provisioned_colour,
-            node_information["memory"]["provisioned"],
-            ansiprint.end(),
-        )
-    )
-
-    # Join it all together
-    ainformation.append("")
-    return "\n".join(ainformation)
-
-
-def format_list(node_list, raw):
-    if raw:
-        ainformation = list()
-        for node in sorted(item["name"] for item in node_list):
-            ainformation.append(node)
-        return "\n".join(ainformation)
-
-    node_list_output = []
-
-    # Determine optimal column widths
-    node_name_length = 5
-    pvc_version_length = 8
-    health_length = 7
-    daemon_state_length = 7
-    coordinator_state_length = 12
-    domain_state_length = 7
-    domains_count_length = 4
-    cpu_count_length = 6
-    load_length = 5
-    mem_total_length = 6
-    mem_used_length = 5
-    mem_free_length = 5
-    mem_alloc_length = 6
-    mem_prov_length = 5
-    for node_information in node_list:
-        # node_name column
-        _node_name_length = len(node_information["name"]) + 1
-        if _node_name_length > node_name_length:
-            node_name_length = _node_name_length
-        # node_pvc_version column
-        _pvc_version_length = len(node_information.get("pvc_version", "N/A")) + 1
-        if _pvc_version_length > pvc_version_length:
-            pvc_version_length = _pvc_version_length
-        # node_health column
-        node_health = node_information.get("health", "N/A")
-        if isinstance(node_health, int):
-            node_health_text = f"{node_health}%"
-        else:
-            node_health_text = node_health
-        _health_length = len(node_health_text) + 1
-        if _health_length > health_length:
-            health_length = _health_length
-        # daemon_state column
-        _daemon_state_length = len(node_information["daemon_state"]) + 1
-        if _daemon_state_length > daemon_state_length:
-            daemon_state_length = _daemon_state_length
-        # coordinator_state column
-        _coordinator_state_length = len(node_information["coordinator_state"]) + 1
-        if _coordinator_state_length > coordinator_state_length:
-            coordinator_state_length = _coordinator_state_length
-        # domain_state column
-        _domain_state_length = len(node_information["domain_state"]) + 1
-        if _domain_state_length > domain_state_length:
-            domain_state_length = _domain_state_length
-        # domains_count column
-        _domains_count_length = len(str(node_information["domains_count"])) + 1
-        if _domains_count_length > domains_count_length:
-            domains_count_length = _domains_count_length
-        # cpu_count column
-        _cpu_count_length = len(str(node_information["cpu_count"])) + 1
-        if _cpu_count_length > cpu_count_length:
-            cpu_count_length = _cpu_count_length
-        # load column
-        _load_length = len(str(node_information["load"])) + 1
-        if _load_length > load_length:
-            load_length = _load_length
-        # mem_total column
-        _mem_total_length = len(str(node_information["memory"]["total"])) + 1
-        if _mem_total_length > mem_total_length:
-            mem_total_length = _mem_total_length
-        # mem_used column
-        _mem_used_length = len(str(node_information["memory"]["used"])) + 1
-        if _mem_used_length > mem_used_length:
-            mem_used_length = _mem_used_length
-        # mem_free column
-        _mem_free_length = len(str(node_information["memory"]["free"])) + 1
-        if _mem_free_length > mem_free_length:
-            mem_free_length = _mem_free_length
-        # mem_alloc column
-        _mem_alloc_length = len(str(node_information["memory"]["allocated"])) + 1
-        if _mem_alloc_length > mem_alloc_length:
-            mem_alloc_length = _mem_alloc_length
-
-        # mem_prov column
-        _mem_prov_length = len(str(node_information["memory"]["provisioned"])) + 1
-        if _mem_prov_length > mem_prov_length:
-            mem_prov_length = _mem_prov_length
-
-    # Format the string (header)
-    node_list_output.append(
-        "{bold}{node_header: <{node_header_length}} {state_header: <{state_header_length}} {resource_header: <{resource_header_length}} {memory_header: <{memory_header_length}}{end_bold}".format(
-            node_header_length=node_name_length
-            + pvc_version_length
-            + health_length
-            + 2,
-            state_header_length=daemon_state_length
-            + coordinator_state_length
-            + domain_state_length
-            + 2,
-            resource_header_length=domains_count_length
-            + cpu_count_length
-            + load_length
-            + 2,
-            memory_header_length=mem_total_length
-            + mem_used_length
-            + mem_free_length
-            + mem_alloc_length
-            + mem_prov_length
-            + 4,
-            bold=ansiprint.bold(),
-            end_bold=ansiprint.end(),
-            node_header="Nodes "
-            + "".join(
-                [
-                    "-"
-                    for _ in range(
-                        6, node_name_length + pvc_version_length + health_length + 1
-                    )
-                ]
-            ),
-            state_header="States "
-            + "".join(
-                [
-                    "-"
-                    for _ in range(
-                        7,
-                        daemon_state_length
-                        + coordinator_state_length
-                        + domain_state_length
-                        + 1,
-                    )
-                ]
-            ),
-            resource_header="Resources "
-            + "".join(
-                [
-                    "-"
-                    for _ in range(
-                        10, domains_count_length + cpu_count_length + load_length + 1
-                    )
-                ]
-            ),
-            memory_header="Memory (M) "
-            + "".join(
-                [
-                    "-"
-                    for _ in range(
-                        11,
-                        mem_total_length
-                        + mem_used_length
-                        + mem_free_length
-                        + mem_alloc_length
-                        + mem_prov_length
-                        + 3,
-                    )
-                ]
-            ),
-        )
-    )
-
-    node_list_output.append(
-        "{bold}{node_name: <{node_name_length}} {node_pvc_version: <{pvc_version_length}} {node_health: <{health_length}} \
-{daemon_state_colour}{node_daemon_state: <{daemon_state_length}}{end_colour} {coordinator_state_colour}{node_coordinator_state: <{coordinator_state_length}}{end_colour} {domain_state_colour}{node_domain_state: <{domain_state_length}}{end_colour} \
-{node_domains_count: <{domains_count_length}} {node_cpu_count: <{cpu_count_length}} {node_load: <{load_length}} \
-{node_mem_total: <{mem_total_length}} {node_mem_used: <{mem_used_length}} {node_mem_free: <{mem_free_length}} {node_mem_allocated: <{mem_alloc_length}} {node_mem_provisioned: <{mem_prov_length}}{end_bold}".format(
-            node_name_length=node_name_length,
-            pvc_version_length=pvc_version_length,
-            health_length=health_length,
-            daemon_state_length=daemon_state_length,
-            coordinator_state_length=coordinator_state_length,
-            domain_state_length=domain_state_length,
-            domains_count_length=domains_count_length,
-            cpu_count_length=cpu_count_length,
-            load_length=load_length,
-            mem_total_length=mem_total_length,
-            mem_used_length=mem_used_length,
-            mem_free_length=mem_free_length,
-            mem_alloc_length=mem_alloc_length,
-            mem_prov_length=mem_prov_length,
-            bold=ansiprint.bold(),
-            end_bold=ansiprint.end(),
-            daemon_state_colour="",
-            coordinator_state_colour="",
-            domain_state_colour="",
-            end_colour="",
-            node_name="Name",
-            node_pvc_version="Version",
-            node_health="Health",
-            node_daemon_state="Daemon",
-            node_coordinator_state="Coordinator",
-            node_domain_state="Domain",
-            node_domains_count="VMs",
-            node_cpu_count="vCPUs",
-            node_load="Load",
-            node_mem_total="Total",
-            node_mem_used="Used",
-            node_mem_free="Free",
-            node_mem_allocated="Alloc",
-            node_mem_provisioned="Prov",
-        )
-    )
-
-    # Format the string (elements)
-    for node_information in sorted(node_list, key=lambda n: n["name"]):
-        (
-            health_colour,
-            daemon_state_colour,
-            coordinator_state_colour,
-            domain_state_colour,
-            mem_allocated_colour,
-            mem_provisioned_colour,
-        ) = getOutputColours(node_information)
-
-        node_health = node_information.get("health", "N/A")
-        if isinstance(node_health, int):
-            node_health_text = f"{node_health}%"
-        else:
-            node_health_text = node_health
-
-        node_list_output.append(
-            "{bold}{node_name: <{node_name_length}} {node_pvc_version: <{pvc_version_length}} {health_colour}{node_health: <{health_length}}{end_colour} \
-{daemon_state_colour}{node_daemon_state: <{daemon_state_length}}{end_colour} {coordinator_state_colour}{node_coordinator_state: <{coordinator_state_length}}{end_colour} {domain_state_colour}{node_domain_state: <{domain_state_length}}{end_colour} \
-{node_domains_count: <{domains_count_length}} {node_cpu_count: <{cpu_count_length}} {node_load: <{load_length}} \
-{node_mem_total: <{mem_total_length}} {node_mem_used: <{mem_used_length}} {node_mem_free: <{mem_free_length}} {mem_allocated_colour}{node_mem_allocated: <{mem_alloc_length}}{end_colour} {mem_provisioned_colour}{node_mem_provisioned: <{mem_prov_length}}{end_colour}{end_bold}".format(
-                node_name_length=node_name_length,
-                pvc_version_length=pvc_version_length,
-                health_length=health_length,
-                daemon_state_length=daemon_state_length,
-                coordinator_state_length=coordinator_state_length,
-                domain_state_length=domain_state_length,
-                domains_count_length=domains_count_length,
-                cpu_count_length=cpu_count_length,
-                load_length=load_length,
-                mem_total_length=mem_total_length,
-                mem_used_length=mem_used_length,
-                mem_free_length=mem_free_length,
-                mem_alloc_length=mem_alloc_length,
-                mem_prov_length=mem_prov_length,
-                bold="",
-                end_bold="",
-                health_colour=health_colour,
-                daemon_state_colour=daemon_state_colour,
-                coordinator_state_colour=coordinator_state_colour,
-                domain_state_colour=domain_state_colour,
-                mem_allocated_colour=mem_allocated_colour,
-                mem_provisioned_colour=mem_allocated_colour,
-                end_colour=ansiprint.end(),
-                node_name=node_information["name"],
-                node_pvc_version=node_information.get("pvc_version", "N/A"),
-                node_health=node_health_text,
-                node_daemon_state=node_information["daemon_state"],
-                node_coordinator_state=node_information["coordinator_state"],
-                node_domain_state=node_information["domain_state"],
-                node_domains_count=node_information["domains_count"],
-                node_cpu_count=node_information["vcpu"]["allocated"],
-                node_load=node_information["load"],
-                node_mem_total=node_information["memory"]["total"],
-                node_mem_used=node_information["memory"]["used"],
-                node_mem_free=node_information["memory"]["free"],
-                node_mem_allocated=node_information["memory"]["allocated"],
-                node_mem_provisioned=node_information["memory"]["provisioned"],
-            )
-        )
-
-    return "\n".join(node_list_output)

client-cli-old/pvc/lib/zkhandler.py

@@ -1,102 +0,0 @@
-#!/usr/bin/env python3
-
-# zkhandler.py - Secure versioned ZooKeeper updates
-# Part of the Parallel Virtual Cluster (PVC) system
-#
-#    Copyright (C) 2018-2022 Joshua M. Boniface <joshua@boniface.me>
-#
-#    This program is free software: you can redistribute it and/or modify
-#    it under the terms of the GNU General Public License as published by
-#    the Free Software Foundation, version 3.
-#
-#    This program is distributed in the hope that it will be useful,
-#    but WITHOUT ANY WARRANTY; without even the implied warranty of
-#    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
-#    GNU General Public License for more details.
-#
-#    You should have received a copy of the GNU General Public License
-#    along with this program.  If not, see <https://www.gnu.org/licenses/>.
-#
-###############################################################################
-
-import uuid
-
-
-# Exists function
-def exists(zk_conn, key):
-    stat = zk_conn.exists(key)
-    if stat:
-        return True
-    else:
-        return False
-
-
-# Child list function
-def listchildren(zk_conn, key):
-    children = zk_conn.get_children(key)
-    return children
-
-
-# Delete key function
-def deletekey(zk_conn, key, recursive=True):
-    zk_conn.delete(key, recursive=recursive)
-
-
-# Data read function
-def readdata(zk_conn, key):
-    data_raw = zk_conn.get(key)
-    data = data_raw[0].decode("utf8")
-    return data
-
-
-# Data write function
-def writedata(zk_conn, kv):
-    # Start up a transaction
-    zk_transaction = zk_conn.transaction()
-
-    # Proceed one KV pair at a time
-    for key in sorted(kv):
-        data = kv[key]
-
-        # Check if this key already exists or not
-        if not zk_conn.exists(key):
-            # We're creating a new key
-            zk_transaction.create(key, str(data).encode("utf8"))
-        else:
-            # We're updating a key with version validation
-            orig_data = zk_conn.get(key)
-            version = orig_data[1].version
-
-            # Set what we expect the new version to be
-            new_version = version + 1
-
-            # Update the data
-            zk_transaction.set_data(key, str(data).encode("utf8"))
-
-            # Set up the check
-            try:
-                zk_transaction.check(key, new_version)
-            except TypeError:
-                print('Zookeeper key "{}" does not match expected version'.format(key))
-                return False
-
-    # Commit the transaction
-    try:
-        zk_transaction.commit()
-        return True
-    except Exception:
-        return False
-
-
-# Write lock function
-def writelock(zk_conn, key):
-    lock_id = str(uuid.uuid1())
-    lock = zk_conn.WriteLock("{}".format(key), lock_id)
-    return lock
-
-
-# Read lock function
-def readlock(zk_conn, key):
-    lock_id = str(uuid.uuid1())
-    lock = zk_conn.ReadLock("{}".format(key), lock_id)
-    return lock

client-cli-old/scripts/README

@@ -1,32 +0,0 @@
-# PVC helper scripts
-
-These helper scripts are included with the PVC client to aid administrators in some meta-functions.
-
-The following scripts are provided for use:
-
-## `migrate_vm`
-
-Migrates a VM, with downtime, from one PVC cluster to another.
-
-`migrate_vm <vm> <source_cluster> <destination_cluster>`
-
-### Arguments
-
- * `vm`: The virtual machine to migrate
- * `source_cluster`: The source PVC cluster; must be a valid cluster to the local PVC client
- * `destination_cluster`: The destination PVC cluster; must be a valid cluster to the local PVC client
-
-## `import_vm`
-
-Imports a VM from another platform into a PVC cluster.
-
-## `export_vm`
-
-Exports a (stopped) VM from a PVC cluster to another platform.
-
-`export_vm <vm> <source_cluster>`
-
-### Arguments
-
- * `vm`: The virtual machine to migrate
- * `source_cluster`: The source PVC cluster; must be a valid cluster to the local PVC client

client-cli-old/scripts/export_vm

@@ -1,98 +0,0 @@
-#!/usr/bin/env bash
-
-# export_vm - Exports a VM from a PVC cluster to local files
-# Part of the Parallel Virtual Cluster (PVC) system
-#
-#    Copyright (C) 2018-2022 Joshua M. Boniface <joshua@boniface.me>
-#
-#    This program is free software: you can redistribute it and/or modify
-#    it under the terms of the GNU General Public License as published by
-#    the Free Software Foundation, version 3.
-#
-#    This program is distributed in the hope that it will be useful,
-#    but WITHOUT ANY WARRANTY; without even the implied warranty of
-#    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
-#    GNU General Public License for more details.
-#
-#    You should have received a copy of the GNU General Public License
-#    along with this program.  If not, see <https://www.gnu.org/licenses/>.
-#
-###############################################################################
-
-set -o errexit
-set -o pipefail
-
-usage() {
-    echo -e "Export a VM from a PVC cluster to local files."
-    echo -e "Usage:"
-    echo -e "  $0 <vm> <source_cluster> [<destination_directory>]"
-    echo -e ""
-    echo -e "Important information:"
-    echo -e " * The local user must have valid SSH access to the primary coordinator in the source_cluster."
-    echo -e " * The user on the cluster primary coordinator must have 'sudo' access."
-    echo -e " * If the VM is not in 'stop' state, it will be shut down."
-    echo -e " * Do not switch the cluster primary coordinator while the script is running."
-    echo -e " * Ensure you have enough space in <destination_directory> to store all VM disk images."
-}
-
-fail() {
-    echo -e "$@"
-    exit 1
-}
-
-# Arguments
-if [[ -z ${1} || -z ${2} ]]; then
-    usage
-    exit 1
-fi
-source_vm="${1}"
-source_cluster="${2}"
-if [[ -n "${3}" ]]; then
-    destination_directory="${3}"
-else
-    destination_directory="."
-fi
-
-# Verify the cluster is reachable
-pvc -c ${source_cluster} status &>/dev/null || fail "Specified source_cluster is not accessible"
-
-# Determine the connection IP
-cluster_address="$( pvc cluster list 2>/dev/null | grep -i "^${source_cluster}" | awk '{ print $2 }' )"
-
-# Attempt to connect to the cluster address
-ssh ${cluster_address} which pvc &>/dev/null || fail "Could not SSH to source_cluster primary coordinator host"
-
-# Verify that the VM exists
-pvc -c ${source_cluster} vm info ${source_vm} &>/dev/null || fail "Specified VM is not present on the cluster"
-
-echo "Verification complete."
-
-# Shut down the VM
-echo -n "Shutting down VM..."
-set +o errexit
-pvc -c ${source_cluster} vm shutdown ${source_vm} &>/dev/null
-shutdown_success=$?
-while ! pvc -c ${source_cluster} vm info ${source_vm} 2>/dev/null | grep '^State' | grep -q -E 'stop|disable'; do
-    sleep 1
-    echo -n "."
-done
-set -o errexit
-echo " done."
-
-# Dump the XML file
-echo -n "Exporting VM configuration file... "
-pvc -c ${source_cluster} vm dump ${source_vm} 1> ${destination_directory}/${source_vm}.xml 2>/dev/null
-echo "done".
-
-# Determine the list of volumes in this VM
-volume_list="$( pvc -c ${source_cluster} vm info --long ${source_vm} 2>/dev/null | grep -w 'rbd' | awk '{ print $3 }' )"
-for volume in ${volume_list}; do
-    volume_pool="$( awk -F '/' '{ print $1 }' <<<"${volume}" )"
-    volume_name="$( awk -F '/' '{ print $2 }' <<<"${volume}" )"
-    volume_size="$( pvc -c ${source_cluster} storage volume list -p ${volume_pool} ${volume_name} 2>/dev/null | grep "^${volume_name}" | awk '{ print $3 }' )"
-    echo -n "Exporting disk ${volume_name} (${volume_size})... "
-    ssh ${cluster_address} sudo rbd map ${volume_pool}/${volume_name} &>/dev/null || fail "Failed to map volume ${volume}"
-    ssh ${cluster_address} sudo dd if="/dev/rbd/${volume_pool}/${volume_name}" bs=1M 2>/dev/null | dd bs=1M of="${destination_directory}/${volume_name}.img" 2>/dev/null
-    ssh ${cluster_address} sudo rbd unmap ${volume_pool}/${volume_name} &>/dev/null || fail "Failed to unmap volume ${volume}"
-    echo "done."
-done

client-cli-old/scripts/force_single_node

@@ -1,118 +0,0 @@
-#!/usr/bin/env bash
-
-# force_single_node - Manually promote a single coordinator node from a degraded cluster
-# Part of the Parallel Virtual Cluster (PVC) system
-#
-#    Copyright (C) 2018-2022 Joshua M. Boniface <joshua@boniface.me>
-#
-#    This program is free software: you can redistribute it and/or modify
-#    it under the terms of the GNU General Public License as published by
-#    the Free Software Foundation, version 3.
-#
-#    This program is distributed in the hope that it will be useful,
-#    but WITHOUT ANY WARRANTY; without even the implied warranty of
-#    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
-#    GNU General Public License for more details.
-#
-#    You should have received a copy of the GNU General Public License
-#    along with this program.  If not, see <https://www.gnu.org/licenses/>.
-#
-###############################################################################
-
-set -o errexit
-set -o pipefail
-
-usage() {
-    echo -e "Manually promote a single coordinator node from a degraded cluster"
-    echo -e ""
-    echo -e "DANGER: This action will cause a permanent split-brain within the cluster"
-    echo -e "        which will have to be corrected manually upon cluster restoration."
-    echo -e ""
-    echo -e "This script is primarily designed for small clusters in situations where 2"
-    echo -e "of the 3 coordinators have become unreachable or shut down. It will promote"
-    echo -e "the remaining lone_node to act as a standalone coordinator, allowing basic"
-    echo -e "cluster functionality to continue in a heavily degraded state until the"
-    echo -e "situation can be rectified. This should only be done in exceptional cases"
-    echo -e "as a disaster recovery mechanism when the remaining nodes will remain down"
-    echo -e "for a significant amount of time but some VMs are required to run. In general,"
-    echo -e "use of this script is not advisable."
-    echo -e ""
-    echo -e "Usage:"
-    echo -e "  $0 <target_cluster> <lone_node>"
-    echo -e ""
-    echo -e "Important information:"
-    echo -e " * The lone_node must be a fully-qualified name that is directly reachable from"
-    echo -e "   the local system via SSH."
-    echo -e " * The local user must have valid SSH access to the lone_node in the cluster."
-    echo -e " * The user on the cluster node must have 'sudo' access."
-}
-
-fail() {
-    echo -e "$@"
-    exit 1
-}
-
-# Arguments
-if [[ -z ${1} || -z ${2} ]]; then
-    usage
-    exit 1
-fi
-target_cluster="${1}"
-lone_node="${2}"
-lone_node_shortname="${lone_node%%.*}"
-
-# Attempt to connect to the node
-ssh ${lone_node} which pvc &>/dev/null || fail "Could not SSH to the lone_node host"
-
-echo "Verification complete."
-
-echo -n "Allowing Ceph single-node operation... "
-temp_monmap="$( ssh ${lone_node} mktemp )"
-ssh ${lone_node} "sudo systemctl stop ceph-mon@${lone_node_shortname}" &>/dev/null
-ssh ${lone_node} "ceph-mon -i ${lone_node_shortname} --extract-monmap ${temp_monmap}" &>/dev/null
-ssh ${lone_node} "sudo cp ${tmp_monmap} /etc/ceph/monmap.orig" &>/dev/null
-mon_list="$( ssh ${lone_node} strings ${temp_monmap} | sort | uniq )"
-for mon in ${mon_list}; do
-    if [[ ${mon} == ${lone_node_shortname} ]]; then
-        continue
-    fi
-    ssh ${lone_node} "sudo monmaptool ${temp_monmap} --rm ${mon}" &>/dev/null
-done
-ssh ${lone_node} "sudo ceph-mon -i ${lone_node_shortname} --inject-monmap ${temp_monmap}" &>/dev/null
-ssh ${lone_node} "sudo systemctl start ceph-mon@${lone_node_shortname}" &>/dev/null
-sleep 5
-ssh ${lone_node} "sudo ceph osd set noout" &>/dev/null
-echo "done."
-echo -e "Restoration steps:"
-echo -e "  sudo systemctl stop ceph-mon@${lone_node_shortname}"
-echo -e "  sudo ceph-mon -i ${lone_node_shortname} --inject-monmap /etc/ceph/monmap.orig"
-echo -e "  sudo systemctl start ceph-mon@${lone_node_shortname}"
-echo -e "  sudo ceph osd unset noout"
-
-echo -n "Allowing Zookeeper single-node operation... "
-temp_zoocfg="$( ssh ${lone_node} mktemp )"
-ssh ${lone_node} "sudo systemctl stop zookeeper"
-ssh ${lone_node} "sudo awk -v lone_node=${lone_node_shortname} '{
-FS="=|:"
-if ( $1 ~ /^server/ ){
-    if ($2 == lone_node) {
-        print $0
-    } else {
-        print "#" $0
-    }
-} else {
-    print $0
-}
-}' /etc/zookeeper/conf/zoo.cfg > ${temp_zoocfg}"
-ssh ${lone_node} "sudo mv /etc/zookeeper/conf/zoo.cfg /etc/zookeeper/conf/zoo.cfg.orig"
-ssh ${lone_node} "sudo mv ${temp_zoocfg} /etc/zookeeper/conf/zoo.cfg"
-ssh ${lone_node} "sudo systemctl start zookeeper"
-echo "done."
-echo -e "Restoration steps:"
-echo -e "  sudo systemctl stop zookeeper"
-echo -e "  sudo mv /etc/zookeeper/conf/zoo.cfg.orig /etc/zookeeper/conf/zoo.cfg"
-echo -e "  sudo systemctl start zookeeper"
-ssh ${lone_node} "sudo systemctl stop ceph-mon@${lone_node_shortname}"
-
-echo ""
-ssh ${lone_node} "sudo pvc status 2>/dev/null"

client-cli-old/scripts/import_vm

@@ -1,80 +0,0 @@
-#!/usr/bin/env bash
-
-# import_vm - Imports a VM to a PVC cluster from local files
-# Part of the Parallel Virtual Cluster (PVC) system
-#
-#    Copyright (C) 2018-2022 Joshua M. Boniface <joshua@boniface.me>
-#
-#    This program is free software: you can redistribute it and/or modify
-#    it under the terms of the GNU General Public License as published by
-#    the Free Software Foundation, version 3.
-#
-#    This program is distributed in the hope that it will be useful,
-#    but WITHOUT ANY WARRANTY; without even the implied warranty of
-#    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
-#    GNU General Public License for more details.
-#
-#    You should have received a copy of the GNU General Public License
-#    along with this program.  If not, see <https://www.gnu.org/licenses/>.
-#
-###############################################################################
-
-set -o errexit
-set -o pipefail
-
-usage() {
-    echo -e "Import a VM to a PVC cluster from local files."
-    echo -e "Usage:"
-    echo -e "  $0 <destination_cluster> <destination_pool> <vm_configuration_file> <vm_disk_file_1> [<vm_disk_file_2>] [...]"
-    echo -e ""
-    echo -e "Important information:"
-    echo -e " * At least one disk must be specified; all disks that are present in vm_configuration_file"
-    echo -e "   should be specified, though this is not strictly requireda."
-    echo -e " * Do not switch the cluster primary coordinator while the script is running."
-    echo -e " * Ensure you have enough space on the destination cluster to store all VM disks."
-}
-
-fail() {
-    echo -e "$@"
-    exit 1
-}
-
-# Arguments
-if [[ -z ${1} || -z ${2} || -z ${3} || -z ${4} ]]; then
-    usage
-    exit 1
-fi
-destination_cluster="${1}"; shift
-destination_pool="${1}"; shift
-vm_config_file="${1}"; shift
-vm_disk_files=( ${@} )
-
-# Verify the cluster is reachable
-pvc -c ${destination_cluster} status &>/dev/null || fail "Specified destination_cluster is not accessible"
-
-# Determine the connection IP
-cluster_address="$( pvc cluster list 2>/dev/null | grep -i "^${destination_cluster}" | awk '{ print $2 }' )"
-
-echo "Verification complete."
-
-# Determine information about the VM from the config file
-parse_xml_field() {
-    field="${1}"
-    line="$( grep -F "<${field}>" ${vm_config_file} )"
-    awk -F '>|<' '{ print $3 }' <<<"${line}"
-}
-vm_name="$( parse_xml_field name )"
-echo "Importing VM ${vm_name}..."
-pvc -c ${destination_cluster} vm define ${vm_config_file} 2>/dev/null
-
-# Create the disks on the cluster
-for disk_file in ${vm_disk_files[@]}; do
-    disk_file_basename="$( basename ${disk_file} )"
-    disk_file_ext="${disk_file_basename##*.}"
-    disk_file_name="$( basename ${disk_file_basename} .${disk_file_ext} )"
-    disk_file_size="$( stat --format="%s" ${disk_file} )"
-
-    echo "Importing disk ${disk_file_name}... "
-    pvc -c ${destination_cluster} storage volume add ${destination_pool} ${disk_file_name} ${disk_file_size}B 2>/dev/null
-    pvc -c ${destination_cluster} storage volume upload ${destination_pool} ${disk_file_name} ${disk_file} 2>/dev/null
-done

client-cli-old/scripts/migrate_vm

@@ -1,115 +0,0 @@
-#!/usr/bin/env bash
-
-# migrate_vm - Exports a VM from a PVC cluster to another PVC cluster
-# Part of the Parallel Virtual Cluster (PVC) system
-#
-#    Copyright (C) 2018-2022 Joshua M. Boniface <joshua@boniface.me>
-#
-#    This program is free software: you can redistribute it and/or modify
-#    it under the terms of the GNU General Public License as published by
-#    the Free Software Foundation, version 3.
-#
-#    This program is distributed in the hope that it will be useful,
-#    but WITHOUT ANY WARRANTY; without even the implied warranty of
-#    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
-#    GNU General Public License for more details.
-#
-#    You should have received a copy of the GNU General Public License
-#    along with this program.  If not, see <https://www.gnu.org/licenses/>.
-#
-###############################################################################
-
-set -o errexit
-set -o pipefail
-
-usage() {
-    echo -e "Export a VM from a PVC cluster to another PVC cluster."
-    echo -e "Usage:"
-    echo -e "  $0 <vm> <source_cluster> <destination_cluster> <destination_pool>"
-    echo -e ""
-    echo -e "Important information:"
-    echo -e " * The local user must have valid SSH access to the primary coordinator in the source_cluster."
-    echo -e " * The user on the cluster primary coordinator must have 'sudo' access."
-    echo -e " * If the VM is not in 'stop' state, it will be shut down."
-    echo -e " * Do not switch the cluster primary coordinator on either cluster while the script is running."
-    echo -e " * Ensure you have enough space on the target cluster to store all VM disks."
-}
-
-fail() {
-    echo -e "$@"
-    exit 1
-}
-
-# Arguments
-if [[ -z ${1} || -z ${2} || -z ${3} || -z ${4} ]]; then
-    usage
-    exit 1
-fi
-source_vm="${1}"
-source_cluster="${2}"
-destination_cluster="${3}"
-destination_pool="${4}"
-
-# Verify each cluster is reachable
-pvc -c ${source_cluster} status &>/dev/null || fail "Specified source_cluster is not accessible"
-pvc -c ${destination_cluster} status &>/dev/null || fail "Specified destination_cluster is not accessible"
-
-# Determine the connection IPs
-source_cluster_address="$( pvc cluster list 2>/dev/null | grep -i "^${source_cluster}" | awk '{ print $2 }' )"
-destination_cluster_address="$( pvc cluster list 2>/dev/null | grep -i "^${destination_cluster}" | awk '{ print $2 }' )"
-
-# Attempt to connect to the cluster addresses
-ssh ${source_cluster_address} which pvc &>/dev/null || fail "Could not SSH to source_cluster primary coordinator host"
-ssh ${destination_cluster_address} which pvc &>/dev/null || fail "Could not SSH to destination_cluster primary coordinator host"
-
-# Verify that the VM exists
-pvc -c ${source_cluster} vm info ${source_vm} &>/dev/null || fail "Specified VM is not present on the source cluster"
-
-echo "Verification complete."
-
-# Shut down the VM
-echo -n "Shutting down VM..."
-set +o errexit
-pvc -c ${source_cluster} vm shutdown ${source_vm} &>/dev/null
-shutdown_success=$?
-while ! pvc -c ${source_cluster} vm info ${source_vm} 2>/dev/null | grep '^State' | grep -q -E 'stop|disable'; do
-    sleep 1
-    echo -n "."
-done
-set -o errexit
-echo " done."
-
-tempfile="$( mktemp )"
-
-# Dump the XML file
-echo -n "Exporting VM configuration file from source cluster... "
-pvc -c ${source_cluster} vm dump ${source_vm} 1> ${tempfile} 2>/dev/null
-echo "done."
-
-# Import the XML file
-echo -n "Importing VM configuration file to destination cluster... "
-pvc -c ${destination_cluster} vm define ${tempfile}
-echo "done."
-
-rm -f ${tempfile}
-
-# Determine the list of volumes in this VM
-volume_list="$( pvc -c ${source_cluster} vm info --long ${source_vm} 2>/dev/null | grep -w 'rbd' | awk '{ print $3 }' )"
-
-# Parse and migrate each volume
-for volume in ${volume_list}; do
-    volume_pool="$( awk -F '/' '{ print $1 }' <<<"${volume}" )"
-    volume_name="$( awk -F '/' '{ print $2 }' <<<"${volume}" )"
-    volume_size="$( pvc -c ${source_cluster} storage volume list -p ${volume_pool} ${volume_name} 2>/dev/null | grep "^${volume_name}" | awk '{ print $3 }' )"
-    echo "Transferring disk ${volume_name} (${volume_size})... "
-    pvc -c ${destination_cluster} storage volume add ${destination_pool} ${volume_name} ${volume_size} 2>/dev/null
-    ssh ${source_cluster_address} sudo rbd map ${volume_pool}/${volume_name} &>/dev/null || fail "Failed to map volume ${volume} on source cluster"
-    ssh ${destination_cluster_address} sudo rbd map ${volume_pool}/${volume_name} &>/dev/null || fail "Failed to map volume ${volume} on destination cluster"
-    ssh ${source_cluster_address} sudo dd if="/dev/rbd/${volume_pool}/${volume_name}" bs=1M 2>/dev/null | pv | ssh ${destination_cluster_address} sudo dd bs=1M of="/dev/rbd/${destination_pool}/${volume_name}" 2>/dev/null
-    ssh ${source_cluster_address} sudo rbd unmap ${volume_pool}/${volume_name} &>/dev/null || fail "Failed to unmap volume ${volume} on source cluster"
-    ssh ${destination_cluster_address} sudo rbd unmap ${volume_pool}/${volume_name} &>/dev/null || fail "Failed to unmap volume ${volume} on destination cluster"
-done
-
-if [[ ${shutdown_success} -eq 0 ]]; then
-    pvc -c ${destination_cluster} vm start ${source_vm}
-fi

client-cli-old/setup.py

@@ -1,20 +0,0 @@
-from setuptools import setup
-
-setup(
-    name="pvc",
-    version="0.9.63",
-    packages=["pvc", "pvc.lib"],
-    install_requires=[
-        "Click",
-        "PyYAML",
-        "lxml",
-        "colorama",
-        "requests",
-        "requests-toolbelt",
-    ],
-    entry_points={
-        "console_scripts": [
-            "pvc = pvc.pvc:cli",
-        ],
-    },
-)

client-cli/autobackup.sample.yaml

@@ -0,0 +1,52 @@
+---
+# Root level configuration key
+autobackup:
+
+  # Backup root path on the node, used as the remote mountpoint
+  # Must be an absolute path beginning with '/'
+  # If remote_mount is enabled, the remote mount will be mounted on this directory
+  # If remote_mount is enabled, it is recommended to use a path under `/tmp` for this
+  # If remote_mount is disabled, a real filesystem must be mounted here (PVC system volumes are small!)
+  backup_root_path: "/tmp/backups"
+
+  # Suffix to the backup root path, used to allow multiple PVC systems to write to a single root path
+  # Must begin with '/'; leave empty to use the backup root path directly
+  # Note that most remote mount options can fake this if needed, but provided to ensure local compatability
+  backup_root_suffix: "/mycluster"
+
+  # VM tag(s) to back up
+  # Only VMs with at least one of the given tag(s) will be backed up; all others will be skipped
+  backup_tags:
+    - "backup"
+    - "mytag"
+
+  # Backup schedule: when and what format to take backups
+  backup_schedule:
+    full_interval:  7  # Number of total backups between full backups; others are incremental
+                       # > If this number is 1, every backup will be a full backup and no incremental
+                       #   backups will be taken
+                       # > If this number is 2, every second backup will be a full backup, etc.
+    full_retention: 2  # Keep this many full backups; the oldest will be deleted when a new one is
+                       # taken, along with all child incremental backups of that backup
+                       # > Should usually be at least 2 when using incrementals (full_interval > 1) to
+                       #   avoid there being too few backups after cleanup from a new full backup
+
+  # Automatic mount settings
+  # These settings permit running an arbitrary set of commands, ideally a "mount" command or similar, to
+  # ensure that a remote filesystem is mounted on the backup root path
+  # While the examples here show absolute paths, that is not required; they will run with the $PATH of the
+  # executing environment (either the "pvc" command on a CLI or a cron/systemd timer)
+  # A "{backup_root_path}" f-string/str.format type variable MAY be present in any cmds string to represent
+  # the above configured root backup path, which is interpolated at runtime
+  # If multiple commands are given, they will be executed in the order given; if no commands are given,
+  # nothing is executed, but the keys MUST be present
+  auto_mount:
+    enabled: no  # Enable automatic mount/unmount support
+    # These commands are executed at the start of the backup run and should mount a filesystem
+    mount_cmds:
+      # This example shows an NFS mount leveraging the backup_root_path variable
+      - "/usr/sbin/mount.nfs -o nfsvers=3 10.0.0.10:/backups {backup_root_path}"
+    # These commands are executed at the end of the backup run and should unmount a filesystem
+    unmount_cmds:
+      # This example shows a generic umount leveraging the backup_root_path variable
+      - "/usr/bin/umount {backup_root_path}"

client-cli/pvc/cli/cli.py

@@ -25,9 +25,8 @@ from functools import wraps
 from json import dump as jdump
 from json import dumps as jdumps
 from json import loads as jloads
-from os import environ, makedirs, path
-from pkg_resources import get_distribution
 from lxml.etree import fromstring, tostring
+from os import environ, makedirs, path
 from re import sub, match
 from yaml import load as yload
 from yaml import SafeLoader as SafeYAMLLoader
@@ -48,34 +47,17 @@ import click
 
 
 ###############################################################################
-# Context and completion handler
+# Context and completion handler, globals
 ###############################################################################
 
 
-CONTEXT_SETTINGS = dict(help_option_names=["-h", "--help"], max_content_width=120)
+CONTEXT_SETTINGS = dict(
+    help_option_names=["-h", "--help"], max_content_width=MAX_CONTENT_WIDTH
+)
 IS_COMPLETION = True if environ.get("_PVC_COMPLETE", "") == "complete" else False
 
 CLI_CONFIG = dict()
 
-if not IS_COMPLETION:
-    cli_client_dir = environ.get("PVC_CLIENT_DIR", None)
-    home_dir = environ.get("HOME", None)
-    if cli_client_dir:
-        store_path = cli_client_dir
-    elif home_dir:
-        store_path = f"{home_dir}/.config/pvc"
-    else:
-        print(
-            "WARNING: No client or home configuration directory found; using /tmp instead"
-        )
-        store_path = "/tmp/pvc"
-
-    if not path.isdir(store_path):
-        makedirs(store_path)
-
-    if not path.isfile(f"{store_path}/{DEFAULT_STORE_FILENAME}"):
-        update_store(store_path, {"local": DEFAULT_STORE_DATA})
-
 
 ###############################################################################
 # Local helper functions
@@ -115,6 +97,8 @@ def version(ctx, param, value):
     if not value or ctx.resilient_parsing:
         return
 
+    from pkg_resources import get_distribution
+
     version = get_distribution("pvc").version
     echo(CLI_CONFIG, f"Parallel Virtual Cluster CLI client version {version}")
     ctx.exit()
@@ -1590,6 +1574,214 @@ def cli_vm_flush_locks(domain):
     finish(retcode, retmsg)
 
 
+###############################################################################
+# > pvc vm backup
+###############################################################################
+@click.group(
+    name="backup",
+    short_help="Manage backups for PVC VMs.",
+    context_settings=CONTEXT_SETTINGS,
+)
+def cli_vm_backup():
+    """
+    Manage backups of VMs in a PVC cluster.
+    """
+    pass
+
+
+###############################################################################
+# > pvc vm backup create
+###############################################################################
+@click.command(name="create", short_help="Create a backup of a virtual machine.")
+@connection_req
+@click.argument("domain")
+@click.argument("backup_path")
+@click.option(
+    "-i",
+    "--incremental",
+    "incremental_parent",
+    default=None,
+    help="Perform an incremental volume backup from this parent backup datestring.",
+)
+@click.option(
+    "-r",
+    "--retain-snapshot",
+    "retain_snapshot",
+    is_flag=True,
+    default=False,
+    help="Retain volume snapshot for future incremental use (full only).",
+)
+def cli_vm_backup_create(domain, backup_path, incremental_parent, retain_snapshot):
+    """
+    Create a backup of virtual machine DOMAIN to BACKUP_PATH on the cluster primary coordinator. DOMAIN may be a UUID or name.
+
+    BACKUP_PATH must be a valid absolute directory path on the cluster "primary" coordinator (see "pvc node list") allowing writes from the API daemon (normally running as "root"). The BACKUP_PATH should be a large storage volume, ideally a remotely mounted filesystem (e.g. NFS, SSHFS, etc.) or non-Ceph-backed disk; PVC does not handle this path, that is up to the administrator to configure and manage.
+
+    The backup will export the VM configuration, metainfo, and a point-in-time snapshot of all attached RBD volumes, using a datestring formatted backup name (i.e. YYYYMMDDHHMMSS).
+
+    The virtual machine DOMAIN may be running, and due to snapshots the backup should be crash-consistent, but will be in an unclean state and this must be considered when restoring from backups.
+
+    Incremental snapshots are possible by specifying the "-i"/"--incremental" option along with a source backup datestring. The snapshots from that source backup must have been retained using the "-r"/"--retain-snapshots" option. Retaining snapshots of incremental backups is not supported as incremental backups cannot be chained.
+
+    Full backup volume images are sparse-allocated, however it is recommended for safety to consider their maximum allocated size when allocated space for the BACKUP_PATH. Incremental volume images are generally small but are dependent entirely on the rate of data change in each volume.
+    """
+
+    echo(
+        CLI_CONFIG,
+        f"Backing up VM '{domain}'... ",
+        newline=False,
+    )
+    retcode, retmsg = pvc.lib.vm.vm_backup(
+        CLI_CONFIG, domain, backup_path, incremental_parent, retain_snapshot
+    )
+    if retcode:
+        echo(CLI_CONFIG, "done.")
+    else:
+        echo(CLI_CONFIG, "failed.")
+    finish(retcode, retmsg)
+
+
+###############################################################################
+# > pvc vm backup restore
+###############################################################################
+@click.command(name="restore", short_help="Restore a backup of a virtual machine.")
+@connection_req
+@click.argument("domain")
+@click.argument("backup_datestring")
+@click.argument("backup_path")
+@click.option(
+    "-r/-R",
+    "--retain-snapshot/--remove-snapshot",
+    "retain_snapshot",
+    is_flag=True,
+    default=True,
+    help="Retain or remove restored (parent, if incremental) snapshot.",
+)
+def cli_vm_backup_restore(domain, backup_datestring, backup_path, retain_snapshot):
+    """
+    Restore the backup BACKUP_DATESTRING of virtual machine DOMAIN stored in BACKUP_PATH on the cluster primary coordinator. DOMAIN may be a UUID or name.
+
+    BACKUP_PATH must be a valid absolute directory path on the cluster "primary" coordinator (see "pvc node list") allowing reads from the API daemon (normally running as "root"). The BACKUP_PATH should be a large storage volume, ideally a remotely mounted filesystem (e.g. NFS, SSHFS, etc.) or non-Ceph-backed disk; PVC does not handle this path, that is up to the administrator to configure and manage.
+
+    The restore will import the VM configuration, metainfo, and the point-in-time snapshot of all attached RBD volumes. Incremental backups will be automatically handled.
+
+    A VM named DOMAIN or with the same UUID must not exist; if a VM with the same name or UUID already exists, it must be removed, or renamed and then undefined (to preserve volumes), before restoring.
+
+    If the "-r"/"--retain-snapshot" option is specified (the default), for incremental restores, only the parent snapshot is kept; for full restores, the restored snapshot is kept. If the "-R"/"--remove-snapshot" option is specified, the imported snapshot is removed.
+
+    WARNING: The "-R"/"--remove-snapshot" option will invalidate any existing incremental backups based on the same incremental parent for the restored VM.
+    """
+
+    echo(
+        CLI_CONFIG,
+        f"Restoring backup {backup_datestring} of VM '{domain}'... ",
+        newline=False,
+    )
+    retcode, retmsg = pvc.lib.vm.vm_restore(
+        CLI_CONFIG, domain, backup_path, backup_datestring, retain_snapshot
+    )
+    if retcode:
+        echo(CLI_CONFIG, "done.")
+    else:
+        echo(CLI_CONFIG, "failed.")
+    finish(retcode, retmsg)
+
+
+###############################################################################
+# > pvc vm backup remove
+###############################################################################
+@click.command(name="remove", short_help="Remove a backup of a virtual machine.")
+@connection_req
+@click.argument("domain")
+@click.argument("backup_datestring")
+@click.argument("backup_path")
+def cli_vm_backup_remove(domain, backup_datestring, backup_path):
+    """
+    Remove the backup BACKUP_DATESTRING, including snapshots, of virtual machine DOMAIN stored in BACKUP_PATH on the cluster primary coordinator. DOMAIN may be a UUID or name.
+
+    WARNING: Removing an incremental parent will invalidate any existing incremental backups based on that backup.
+    """
+
+    echo(
+        CLI_CONFIG,
+        f"Removing backup {backup_datestring} of VM '{domain}'... ",
+        newline=False,
+    )
+    retcode, retmsg = pvc.lib.vm.vm_remove_backup(
+        CLI_CONFIG, domain, backup_path, backup_datestring
+    )
+    if retcode:
+        echo(CLI_CONFIG, "done.")
+    else:
+        echo(CLI_CONFIG, "failed.")
+    finish(retcode, retmsg)
+
+
+###############################################################################
+# > pvc vm autobackup
+###############################################################################
+@click.command(
+    name="autobackup", short_help="Perform automatic virtual machine backups."
+)
+@connection_req
+@click.option(
+    "-f",
+    "--configuration",
+    "autobackup_cfgfile",
+    envvar="PVC_AUTOBACKUP_CFGFILE",
+    default=DEFAULT_AUTOBACKUP_FILENAME,
+    show_default=True,
+    help="Override default config file location.",
+)
+@click.option(
+    "--force-full",
+    "force_full_flag",
+    default=False,
+    is_flag=True,
+    help="Force all backups to be full backups this run.",
+)
+@click.option(
+    "--cron",
+    "cron_flag",
+    default=False,
+    is_flag=True,
+    help="Cron mode; don't error exit if this isn't the primary coordinator.",
+)
+def cli_vm_autobackup(autobackup_cfgfile, force_full_flag, cron_flag):
+    """
+    Perform automated backups of VMs, with integrated cleanup and full/incremental scheduling.
+
+    This command enables automatic backup of PVC VMs at the block level, leveraging the various "pvc vm backup"
+    functions with an internal rentention and cleanup system as well as determination of full vs. incremental
+    backups at different intervals. VMs are selected based on configured VM tags. The destination storage
+    may either be local, or provided by a remote filesystem which is automatically mounted and unmounted during
+    the backup run via a set of configured commands before and after the backup run.
+
+    NOTE: This command performs its tasks in a local context. It MUST be run from the cluster's active primary
+    coordinator using the "local" connection only; if either is not correct, the command will error.
+
+    NOTE: This command should be run as the same user as the API daemon, usually "root" with "sudo -E" or in
+    a cronjob as "root", to ensure permissions are correct on the backup files. Failure to do so will still take
+    the backup, but the state update write will likely fail and the backup will become untracked. The command
+    will prompt for confirmation if it is found not to be running as "root" and this cannot be bypassed.
+
+    This command should be run from cron or a timer at a regular interval (e.g. daily, hourly, etc.) which defines
+    how often backups are taken. Backup format (full/incremental) and retention is based only on the number of
+    recorded backups, not on the time interval between them. Backups taken manually outside of the "autobackup"
+    command are not counted towards the format or retention of autobackups.
+
+    The PVC_AUTOBACKUP_CFGFILE envvar or "-f"/"--configuration" option can be used to override the default
+    configuration file path if required by a particular run. For full details of the possible options, please
+    see the example configuration file at "/usr/share/pvc/autobackup.sample.yaml".
+
+    The "--force-full" option can be used to force all configured VMs to perform a "full" level backup this run,
+    which can help synchronize the backups of existing VMs with new ones.
+    """
+
+    # All work here is done in the helper function for portability; we don't even use "finish"
+    vm_autobackup(CLI_CONFIG, autobackup_cfgfile, force_full_flag, cron_flag)
+
+
 ###############################################################################
 # > pvc vm tag
 ###############################################################################
@@ -3457,14 +3649,14 @@ def cli_storage_pool():
     show_default=True,
     required=False,
     help="""
-    The replication configuration, specifying both a "copies" and "mincopies" value, separated by a comma, e.g. "copies=3,mincopies=2". The "copies" value specifies the total number of replicas and should not exceed the total number of nodes; the "mincopies" value specifies the minimum number of available copies to allow writes. For additional details please see the Cluster Architecture documentation.
+    The replication configuration, specifying both a "copies" and "mincopies" value, separated by a comma, e.g. "copies=3,mincopies=2". The "copies" value specifies the total number of replicas and the "mincopies" value specifies the minimum number of active replicas to allow I/O. For additional details please see the documentation.
     """,
 )
 def cli_storage_pool_add(name, pgs, tier, replcfg):
     """
     Add a new Ceph RBD pool with name NAME and PGS placement groups.
 
-    The placement group count must be a non-zero power of 2.
+    The placement group count must be a non-zero power of 2. Generally you should choose a PGS number such that there will be 50-150 PGs on each OSD in a single node (before replicas); 64, 128, or 256 are good values for small clusters (1-5 OSDs per node); higher values are recommended for higher node or OSD counts. For additional details please see the documentation.
     """
 
     retcode, retmsg = pvc.lib.storage.ceph_pool_add(
@@ -3503,9 +3695,9 @@ def cli_storage_pool_set_pgs(name, pgs):
     """
     Set the placement groups (PGs) count for the pool NAME to PGS.
 
-    The placement group count must be a non-zero power of 2.
+    The placement group count must be a non-zero power of 2. Generally you should choose a PGS number such that there will be 50-150 PGs on each OSD in a single node (before replicas); 64, 128, or 256 are good values for small clusters (1-5 OSDs per node); higher values are recommended for higher node or OSD counts. For additional details please see the documentation.
 
-    Placement group counts may be increased or decreased as required though frequent alteration is not recommended.
+    Placement group counts may be increased or decreased as required though frequent alteration is not recommended. Placement group alterations are intensive operations on the storage cluster.
     """
 
     retcode, retmsg = pvc.lib.storage.ceph_pool_set_pgs(CLI_CONFIG, name, pgs)
@@ -5614,6 +5806,29 @@ def cli(
     """
 
     global CLI_CONFIG
+    CLI_CONFIG["quiet"] = _quiet
+    CLI_CONFIG["silent"] = _silent
+
+    cli_client_dir = environ.get("PVC_CLIENT_DIR", None)
+    home_dir = environ.get("HOME", None)
+    if cli_client_dir:
+        store_path = cli_client_dir
+    elif home_dir:
+        store_path = f"{home_dir}/.config/pvc"
+    else:
+        echo(
+            CLI_CONFIG,
+            "WARNING: No client or home configuration directory found; using /tmp instead",
+            stderr=True,
+        )
+        store_path = "/tmp/pvc"
+
+    if not path.isdir(store_path):
+        makedirs(store_path)
+
+    if not path.isfile(f"{store_path}/{DEFAULT_STORE_FILENAME}"):
+        update_store(store_path, {"local": DEFAULT_STORE_DATA})
+
     store_data = get_store(store_path)
 
     # If the connection isn't in the store, mark it bad but pass the value
@@ -5659,6 +5874,11 @@ cli_vm.add_command(cli_vm_move)
 cli_vm.add_command(cli_vm_migrate)
 cli_vm.add_command(cli_vm_unmigrate)
 cli_vm.add_command(cli_vm_flush_locks)
+cli_vm_backup.add_command(cli_vm_backup_create)
+cli_vm_backup.add_command(cli_vm_backup_restore)
+cli_vm_backup.add_command(cli_vm_backup_remove)
+cli_vm.add_command(cli_vm_backup)
+cli_vm.add_command(cli_vm_autobackup)
 cli_vm_tag.add_command(cli_vm_tag_get)
 cli_vm_tag.add_command(cli_vm_tag_add)
 cli_vm_tag.add_command(cli_vm_tag_remove)

client-cli/pvc/cli/helpers.py

@@ -20,25 +20,33 @@
 ###############################################################################
 
 from click import echo as click_echo
-from click import progressbar
+from click import progressbar, confirm
+from datetime import datetime
 from distutils.util import strtobool
+from getpass import getuser
 from json import load as jload
 from json import dump as jdump
-from os import chmod, environ, getpid, path
+from os import chmod, environ, getpid, path, makedirs
+from re import findall
 from socket import gethostname
+from subprocess import run, PIPE
 from sys import argv
 from syslog import syslog, openlog, closelog, LOG_AUTH
 from time import sleep
 from yaml import load as yload
-from yaml import BaseLoader
+from yaml import BaseLoader, SafeLoader
 
 import pvc.lib.provisioner
+import pvc.lib.vm
+import pvc.lib.node
 
 
 DEFAULT_STORE_DATA = {"cfgfile": "/etc/pvc/pvcapid.yaml"}
 DEFAULT_STORE_FILENAME = "pvc.json"
 DEFAULT_API_PREFIX = "/api/v1"
 DEFAULT_NODE_HOSTNAME = gethostname().split(".")[0]
+DEFAULT_AUTOBACKUP_FILENAME = "/etc/pvc/autobackup.yaml"
+MAX_CONTENT_WIDTH = 120
 
 
 def echo(config, message, newline=True, stderr=False):
@@ -65,10 +73,9 @@ def audit():
     """
 
     args = argv
-    args[0] = "pvc"
     pid = getpid()
 
-    openlog(facility=LOG_AUTH, ident=f"{args[0]}[{pid}]")
+    openlog(facility=LOG_AUTH, ident=f"{args[0].split('/')[-1]}[{pid}]")
     syslog(
         f"""client audit: command "{' '.join(args)}" by user {environ.get('USER', None)}"""
     )
@@ -239,3 +246,322 @@ def wait_for_provisioner(CLI_CONFIG, task_id):
     retdata = task_status.get("state") + ": " + task_status.get("status")
 
     return retdata
+
+
+def get_autobackup_config(CLI_CONFIG, cfgfile):
+    try:
+        config = dict()
+        with open(cfgfile) as fh:
+            backup_config = yload(fh, Loader=SafeLoader)["autobackup"]
+
+        config["backup_root_path"] = backup_config["backup_root_path"]
+        config["backup_root_suffix"] = backup_config["backup_root_suffix"]
+        config["backup_tags"] = backup_config["backup_tags"]
+        config["backup_schedule"] = backup_config["backup_schedule"]
+        config["auto_mount_enabled"] = backup_config["auto_mount"]["enabled"]
+        if config["auto_mount_enabled"]:
+            config["mount_cmds"] = list()
+            _mount_cmds = backup_config["auto_mount"]["mount_cmds"]
+            for _mount_cmd in _mount_cmds:
+                if "{backup_root_path}" in _mount_cmd:
+                    _mount_cmd = _mount_cmd.format(
+                        backup_root_path=backup_config["backup_root_path"]
+                    )
+                config["mount_cmds"].append(_mount_cmd)
+
+            config["unmount_cmds"] = list()
+            _unmount_cmds = backup_config["auto_mount"]["unmount_cmds"]
+            for _unmount_cmd in _unmount_cmds:
+                if "{backup_root_path}" in _unmount_cmd:
+                    _unmount_cmd = _unmount_cmd.format(
+                        backup_root_path=backup_config["backup_root_path"]
+                    )
+                config["unmount_cmds"].append(_unmount_cmd)
+
+    except FileNotFoundError:
+        echo(CLI_CONFIG, "ERROR: Specified backup configuration does not exist!")
+        exit(1)
+    except KeyError as e:
+        echo(CLI_CONFIG, f"ERROR: Backup configuration is invalid: {e}")
+        exit(1)
+
+    return config
+
+
+def vm_autobackup(
+    CLI_CONFIG,
+    autobackup_cfgfile=DEFAULT_AUTOBACKUP_FILENAME,
+    force_full_flag=False,
+    cron_flag=False,
+):
+    """
+    Perform automatic backups of VMs based on an external config file.
+    """
+
+    # Validate that we are running on the current primary coordinator of the 'local' cluster connection
+    real_connection = CLI_CONFIG["connection"]
+    CLI_CONFIG["connection"] = "local"
+    retcode, retdata = pvc.lib.node.node_info(CLI_CONFIG, DEFAULT_NODE_HOSTNAME)
+    if not retcode or retdata.get("coordinator_state") != "primary":
+        if cron_flag:
+            echo(
+                CLI_CONFIG,
+                "Current host is not the primary coordinator of the local cluster and running in cron mode. Exiting cleanly.",
+            )
+            exit(0)
+        else:
+            echo(
+                CLI_CONFIG,
+                f"ERROR: Current host is not the primary coordinator of the local cluster; got connection '{real_connection}', host '{DEFAULT_NODE_HOSTNAME}'.",
+            )
+            echo(
+                CLI_CONFIG,
+                "Autobackup MUST be run from the cluster active primary coordinator using the 'local' connection. See '-h'/'--help' for details.",
+            )
+            exit(1)
+
+    # Ensure we're running as root, or show a warning & confirmation
+    if getuser() != "root":
+        confirm(
+            "WARNING: You are not running this command as 'root'. This command should be run under the same user as the API daemon, which is usually 'root'. Are you sure you want to continue?",
+            prompt_suffix=" ",
+            abort=True,
+        )
+
+    # Load our YAML config
+    autobackup_config = get_autobackup_config(CLI_CONFIG, autobackup_cfgfile)
+
+    # Get a list of all VMs on the cluster
+    # We don't do tag filtering here, because we could match an arbitrary number of tags; instead, we
+    # parse the list after
+    retcode, retdata = pvc.lib.vm.vm_list(CLI_CONFIG, None, None, None, None, None)
+    if not retcode:
+        echo(CLI_CONFIG, f"ERROR: Failed to fetch VM list: {retdata}")
+        exit(1)
+    cluster_vms = retdata
+
+    # Parse the list to match tags; too complex for list comprehension alas
+    backup_vms = list()
+    for vm in cluster_vms:
+        vm_tag_names = [t["name"] for t in vm["tags"]]
+        matching_tags = (
+            True
+            if len(
+                set(vm_tag_names).intersection(set(autobackup_config["backup_tags"]))
+            )
+            > 0
+            else False
+        )
+        if matching_tags:
+            backup_vms.append(vm["name"])
+
+    if len(backup_vms) < 1:
+        echo(CLI_CONFIG, "Found no suitable VMs for autobackup.")
+        exit(0)
+
+    # Pretty print the names of the VMs we'll back up (to stderr)
+    maxnamelen = max([len(n) for n in backup_vms]) + 2
+    cols = 1
+    while (cols * maxnamelen + maxnamelen + 2) <= MAX_CONTENT_WIDTH:
+        cols += 1
+    rows = len(backup_vms) // cols
+    vm_list_rows = list()
+    for row in range(0, rows + 1):
+        row_start = row * cols
+        row_end = (row * cols) + cols
+        row_str = ""
+        for x in range(row_start, row_end):
+            if x < len(backup_vms):
+                row_str += "{:<{}}".format(backup_vms[x], maxnamelen)
+        vm_list_rows.append(row_str)
+
+    echo(CLI_CONFIG, f"Found {len(backup_vms)} suitable VM(s) for autobackup.")
+    echo(CLI_CONFIG, "Full VM list:", stderr=True)
+    echo(CLI_CONFIG, "  {}".format("\n  ".join(vm_list_rows)), stderr=True)
+    echo(CLI_CONFIG, "", stderr=True)
+
+    if autobackup_config["auto_mount_enabled"]:
+        # Execute each mount_cmds command in sequence
+        for cmd in autobackup_config["mount_cmds"]:
+            echo(
+                CLI_CONFIG,
+                f"Executing mount command '{cmd.split()[0]}'... ",
+                newline=False,
+            )
+            tstart = datetime.now()
+            ret = run(
+                cmd.split(),
+                stdout=PIPE,
+                stderr=PIPE,
+            )
+            tend = datetime.now()
+            ttot = tend - tstart
+            if ret.returncode != 0:
+                echo(
+                    CLI_CONFIG,
+                    f"failed. [{ttot.seconds}s]",
+                )
+                echo(
+                    CLI_CONFIG,
+                    f"Exiting; command reports: {ret.stderr.decode().strip()}",
+                )
+                exit(1)
+            else:
+                echo(CLI_CONFIG, f"done. [{ttot.seconds}s]")
+
+    # For each VM, perform the backup
+    for vm in backup_vms:
+        backup_suffixed_path = f"{autobackup_config['backup_root_path']}{autobackup_config['backup_root_suffix']}"
+        if not path.exists(backup_suffixed_path):
+            makedirs(backup_suffixed_path)
+
+        backup_path = f"{backup_suffixed_path}/{vm}"
+        autobackup_state_file = f"{backup_path}/.autobackup.json"
+        if not path.exists(backup_path) or not path.exists(autobackup_state_file):
+            # There are no new backups so the list is empty
+            state_data = dict()
+            tracked_backups = list()
+        else:
+            with open(autobackup_state_file) as fh:
+                state_data = jload(fh)
+            tracked_backups = state_data["tracked_backups"]
+
+        full_interval = autobackup_config["backup_schedule"]["full_interval"]
+        full_retention = autobackup_config["backup_schedule"]["full_retention"]
+
+        full_backups = [b for b in tracked_backups if b["type"] == "full"]
+        if len(full_backups) > 0:
+            last_full_backup = full_backups[0]
+            last_full_backup_idx = tracked_backups.index(last_full_backup)
+            if force_full_flag:
+                this_backup_type = "forced-full"
+                this_backup_incremental_parent = None
+                this_backup_retain_snapshot = True
+            elif last_full_backup_idx >= full_interval - 1:
+                this_backup_type = "full"
+                this_backup_incremental_parent = None
+                this_backup_retain_snapshot = True
+            else:
+                this_backup_type = "incremental"
+                this_backup_incremental_parent = last_full_backup["datestring"]
+                this_backup_retain_snapshot = False
+        else:
+            # The very first backup must be full to start the tree
+            this_backup_type = "full"
+            this_backup_incremental_parent = None
+            this_backup_retain_snapshot = True
+
+        # Perform the backup
+        echo(
+            CLI_CONFIG,
+            f"Backing up VM '{vm}' ({this_backup_type})... ",
+            newline=False,
+        )
+        tstart = datetime.now()
+        retcode, retdata = pvc.lib.vm.vm_backup(
+            CLI_CONFIG,
+            vm,
+            backup_suffixed_path,
+            incremental_parent=this_backup_incremental_parent,
+            retain_snapshot=this_backup_retain_snapshot,
+        )
+        tend = datetime.now()
+        ttot = tend - tstart
+        if not retcode:
+            echo(CLI_CONFIG, f"failed. [{ttot.seconds}s]")
+            echo(CLI_CONFIG, f"Skipping cleanups; command reports: {retdata}")
+            continue
+        else:
+            backup_datestring = findall(r"[0-9]{14}", retdata)[0]
+            echo(
+                CLI_CONFIG,
+                f"done. Backup '{backup_datestring}' created. [{ttot.seconds}s]",
+            )
+
+        # Read backup file to get details
+        backup_json_file = f"{backup_path}/{backup_datestring}/pvcbackup.json"
+        with open(backup_json_file) as fh:
+            backup_json = jload(fh)
+        backup = {
+            "datestring": backup_json["datestring"],
+            "type": backup_json["type"],
+            "parent": backup_json["incremental_parent"],
+            "retained_snapshot": backup_json["retained_snapshot"],
+        }
+        tracked_backups.insert(0, backup)
+
+        # Delete any full backups that are expired
+        marked_for_deletion = list()
+        found_full_count = 0
+        for backup in tracked_backups:
+            if backup["type"] == "full":
+                found_full_count += 1
+                if found_full_count > full_retention:
+                    marked_for_deletion.append(backup)
+
+        # Depete any incremental backups that depend on marked parents
+        for backup in tracked_backups:
+            if backup["type"] == "incremental" and backup["parent"] in [
+                b["datestring"] for b in marked_for_deletion
+            ]:
+                marked_for_deletion.append(backup)
+
+        # Execute deletes
+        for backup_to_delete in marked_for_deletion:
+            echo(
+                CLI_CONFIG,
+                f"Removing old VM '{vm}' backup '{backup_to_delete['datestring']}' ({backup_to_delete['type']})... ",
+                newline=False,
+            )
+            tstart = datetime.now()
+            retcode, retdata = pvc.lib.vm.vm_remove_backup(
+                CLI_CONFIG,
+                vm,
+                backup_suffixed_path,
+                backup_to_delete["datestring"],
+            )
+            tend = datetime.now()
+            ttot = tend - tstart
+            if not retcode:
+                echo(CLI_CONFIG, f"failed. [{ttot.seconds}s]")
+                echo(
+                    CLI_CONFIG,
+                    f"Skipping removal from tracked backups; command reports: {retdata}",
+                )
+                continue
+            else:
+                tracked_backups.remove(backup_to_delete)
+                echo(CLI_CONFIG, f"done. [{ttot.seconds}s]")
+
+        # Update tracked state information
+        state_data["tracked_backups"] = tracked_backups
+        with open(autobackup_state_file, "w") as fh:
+            jdump(state_data, fh)
+
+    if autobackup_config["auto_mount_enabled"]:
+        # Execute each unmount_cmds command in sequence
+        for cmd in autobackup_config["unmount_cmds"]:
+            echo(
+                CLI_CONFIG,
+                f"Executing unmount command '{cmd.split()[0]}'... ",
+                newline=False,
+            )
+            tstart = datetime.now()
+            ret = run(
+                cmd.split(),
+                stdout=PIPE,
+                stderr=PIPE,
+            )
+            tend = datetime.now()
+            ttot = tend - tstart
+            if ret.returncode != 0:
+                echo(
+                    CLI_CONFIG,
+                    f"failed. [{ttot.seconds}s]",
+                )
+                echo(
+                    CLI_CONFIG,
+                    f"Continuing; command reports: {ret.stderr.decode().strip()}",
+                )
+            else:
+                echo(CLI_CONFIG, f"done. [{ttot.seconds}s]")

client-cli/pvc/lib/vm.py

@@ -433,6 +433,70 @@ def vm_locks(config, vm):
     return retstatus, response.json().get("message", "")
 
 
+def vm_backup(config, vm, backup_path, incremental_parent=None, retain_snapshot=False):
+    """
+    Create a backup of {vm} and its volumes to a local primary coordinator filesystem path
+
+    API endpoint: POST /vm/{vm}/backup
+    API arguments: backup_path={backup_path}, incremental_parent={incremental_parent}, retain_snapshot={retain_snapshot}
+    API schema: {"message":"{data}"}
+    """
+    params = {
+        "backup_path": backup_path,
+        "incremental_parent": incremental_parent,
+        "retain_snapshot": retain_snapshot,
+    }
+    response = call_api(config, "post", "/vm/{vm}/backup".format(vm=vm), params=params)
+
+    if response.status_code != 200:
+        return False, response.json().get("message", "")
+    else:
+        return True, response.json().get("message", "")
+
+
+def vm_remove_backup(config, vm, backup_path, backup_datestring):
+    """
+    Remove a backup of {vm}, including snapshots, from a local primary coordinator filesystem path
+
+    API endpoint: DELETE /vm/{vm}/backup
+    API arguments: backup_path={backup_path}, backup_datestring={backup_datestring}
+    API schema: {"message":"{data}"}
+    """
+    params = {
+        "backup_path": backup_path,
+        "backup_datestring": backup_datestring,
+    }
+    response = call_api(
+        config, "delete", "/vm/{vm}/backup".format(vm=vm), params=params
+    )
+
+    if response.status_code != 200:
+        return False, response.json().get("message", "")
+    else:
+        return True, response.json().get("message", "")
+
+
+def vm_restore(config, vm, backup_path, backup_datestring, retain_snapshot=False):
+    """
+    Restore a backup of {vm} and its volumes from a local primary coordinator filesystem path
+
+    API endpoint: POST /vm/{vm}/restore
+    API arguments: backup_path={backup_path}, backup_datestring={backup_datestring}, retain_snapshot={retain_snapshot}
+    API schema: {"message":"{data}"}
+    """
+    params = {
+        "backup_path": backup_path,
+        "backup_datestring": backup_datestring,
+        "retain_snapshot": retain_snapshot,
+    }
+    response = call_api(config, "post", "/vm/{vm}/restore".format(vm=vm), params=params)
+
+    if response.status_code != 200:
+        return False, response.json().get("message", "")
+    else:
+        return True, response.json().get("message", "")
+
+
 def vm_vcpus_set(config, vm, vcpus, topology, restart):
     """
     Set the vCPU count of the VM with topology

client-cli/setup.py

@@ -2,7 +2,7 @@ from setuptools import setup
 
 setup(
     name="pvc",
-    version="0.9.78",
+    version="0.9.80",
     packages=["pvc.cli", "pvc.lib"],
     install_requires=[
         "Click",

daemon-common/ceph.py

@@ -1113,23 +1113,24 @@ def getCephSnapshots(zkhandler, pool, volume):
     return snapshot_list
 
 
-def add_snapshot(zkhandler, pool, volume, name):
+def add_snapshot(zkhandler, pool, volume, name, zk_only=False):
     if not verifyVolume(zkhandler, pool, volume):
         return False, 'ERROR: No volume with name "{}" is present in pool "{}".'.format(
             volume, pool
         )
 
     # 1. Create the snapshot
-    retcode, stdout, stderr = common.run_os_command(
-        "rbd snap create {}/{}@{}".format(pool, volume, name)
-    )
-    if retcode:
-        return (
-            False,
-            'ERROR: Failed to create RBD snapshot "{}" of volume "{}" in pool "{}": {}'.format(
-                name, volume, pool, stderr
-            ),
+    if not zk_only:
+        retcode, stdout, stderr = common.run_os_command(
+            "rbd snap create {}/{}@{}".format(pool, volume, name)
         )
+        if retcode:
+            return (
+                False,
+                'ERROR: Failed to create RBD snapshot "{}" of volume "{}" in pool "{}": {}'.format(
+                    name, volume, pool, stderr
+                ),
+            )
 
     # 2. Add the snapshot to Zookeeper
     zkhandler.write(

daemon-common/common.py

@@ -146,7 +146,11 @@ def run_os_daemon(command_string, environment=None, logfile=None):
 # Run a local OS command via shell
 #
 def run_os_command(command_string, background=False, environment=None, timeout=None):
-    command = shlex_split(command_string)
+    if not isinstance(command_string, list):
+        command = shlex_split(command_string)
+    else:
+        command = command_string
+
     if background:
 
         def runcmd():

daemon-common/vm.py

@@ -21,12 +21,18 @@
 
 import time
 import re
+import os.path
 import lxml.objectify
 import lxml.etree
 
-from distutils.util import strtobool
-from uuid import UUID
 from concurrent.futures import ThreadPoolExecutor
+from datetime import datetime
+from distutils.util import strtobool
+from json import dump as jdump
+from json import load as jload
+from shutil import rmtree
+from socket import gethostname
+from uuid import UUID
 
 import daemon_lib.common as common
 
@@ -1175,13 +1181,15 @@ def get_info(zkhandler, domain):
     return True, domain_information
 
 
-def get_list(zkhandler, node, state, tag, limit, is_fuzzy=True, negate=False):
-    if node:
+def get_list(
+    zkhandler, node=None, state=None, tag=None, limit=None, is_fuzzy=True, negate=False
+):
+    if node is not None:
         # Verify node is valid
         if not common.verifyNode(zkhandler, node):
             return False, 'Specified node "{}" is invalid.'.format(node)
 
-    if state:
+    if state is not None:
         valid_states = [
             "start",
             "restart",
@@ -1200,7 +1208,7 @@ def get_list(zkhandler, node, state, tag, limit, is_fuzzy=True, negate=False):
     full_vm_list.sort()
 
     # Set our limit to a sensible regex
-    if limit:
+    if limit is not None:
         # Check if the limit is a UUID
         is_limit_uuid = False
         try:
@@ -1229,7 +1237,7 @@ def get_list(zkhandler, node, state, tag, limit, is_fuzzy=True, negate=False):
         is_state_match = False
 
         # Check on limit
-        if limit:
+        if limit is not None:
             # Try to match the limit against the UUID (if applicable) and name
             try:
                 if is_limit_uuid and re.fullmatch(limit, vm):
@@ -1241,7 +1249,7 @@ def get_list(zkhandler, node, state, tag, limit, is_fuzzy=True, negate=False):
         else:
             is_limit_match = True
 
-        if tag:
+        if tag is not None:
             vm_tags = zkhandler.children(("domain.meta.tags", vm))
             if negate and tag not in vm_tags:
                 is_tag_match = True
@@ -1251,7 +1259,7 @@ def get_list(zkhandler, node, state, tag, limit, is_fuzzy=True, negate=False):
             is_tag_match = True
 
         # Check on node
-        if node:
+        if node is not None:
             vm_node = zkhandler.read(("domain.node", vm))
             if negate and vm_node != node:
                 is_node_match = True
@@ -1261,7 +1269,7 @@ def get_list(zkhandler, node, state, tag, limit, is_fuzzy=True, negate=False):
             is_node_match = True
 
         # Check on state
-        if state:
+        if state is not None:
             vm_state = zkhandler.read(("domain.state", vm))
             if negate and vm_state != state:
                 is_state_match = True
@@ -1297,3 +1305,541 @@ def get_list(zkhandler, node, state, tag, limit, is_fuzzy=True, negate=False):
                 pass
 
     return True, sorted(vm_data_list, key=lambda d: d["name"])
+
+
+def backup_vm(
+    zkhandler, domain, backup_path, incremental_parent=None, retain_snapshot=False
+):
+
+    tstart = time.time()
+
+    # 0. Validations
+    # Disallow retaining snapshots with an incremental parent
+    if incremental_parent is not None and retain_snapshot:
+        return (
+            False,
+            "ERROR: Retaining snapshots of incremental backups is not supported!",
+        )
+
+    # Validate that VM exists in cluster
+    dom_uuid = getDomainUUID(zkhandler, domain)
+    if not dom_uuid:
+        return False, 'ERROR: Could not find VM "{}" in the cluster!'.format(domain)
+
+    # Validate that the target path is valid
+    if not re.match(r"^/", backup_path):
+        return (
+            False,
+            f"ERROR: Target path {backup_path} is not a valid absolute path on the primary coordinator!",
+        )
+
+    # Ensure that backup_path (on this node) exists
+    if not os.path.isdir(backup_path):
+        return False, f"ERROR: Target path {backup_path} does not exist!"
+
+    # 1. Get information about VM
+    vm_detail = get_list(zkhandler, limit=dom_uuid, is_fuzzy=False)[1][0]
+    if not isinstance(vm_detail, dict):
+        return False, f"ERROR: VM listing returned invalid data: {vm_detail}"
+
+    vm_volumes = list()
+    for disk in vm_detail["disks"]:
+        if disk["type"] != "rbd":
+            continue
+
+        pool, volume = disk["name"].split("/")
+
+        retcode, retdata = ceph.get_list_volume(zkhandler, pool, volume, is_fuzzy=False)
+        if not retcode or len(retdata) != 1:
+            if len(retdata) < 1:
+                retdata = "No volumes returned."
+            elif len(retdata) > 1:
+                retdata = "Multiple volumes returned."
+            return (
+                False,
+                f"ERROR: Failed to get volume details for {pool}/{volume}: {retdata}",
+            )
+
+        try:
+            size = retdata[0]["stats"]["size"]
+        except Exception as e:
+            return False, f"ERROR: Failed to get volume size for {pool}/{volume}: {e}"
+
+        vm_volumes.append((pool, volume, size))
+
+    # 2a. Validate that all volumes exist (they should, but just in case)
+    for pool, volume, _ in vm_volumes:
+        if not ceph.verifyVolume(zkhandler, pool, volume):
+            return (
+                False,
+                f"ERROR: VM defines a volume {pool}/{volume} which does not exist!",
+            )
+
+    # 2b. Validate that, if an incremental_parent is given, it is valid
+    # The incremental parent is just a datestring
+    if incremental_parent is not None:
+        for pool, volume, _ in vm_volumes:
+            if not ceph.verifySnapshot(
+                zkhandler, pool, volume, f"backup_{incremental_parent}"
+            ):
+                return (
+                    False,
+                    f"ERROR: Incremental parent {incremental_parent} given, but no snapshots were found; cannot export an incremental backup.",
+                )
+
+        export_fileext = "rbddiff"
+    else:
+        export_fileext = "rbdimg"
+
+    # 2c. Validate that there's enough space on the target
+    # TODO
+
+    # 3. Set datestring in YYYYMMDDHHMMSS format
+    now = datetime.now()
+    datestring = now.strftime("%Y%m%d%H%M%S")
+
+    snapshot_name = f"backup_{datestring}"
+
+    # 4. Create destination directory
+    vm_target_root = f"{backup_path}/{domain}"
+    vm_target_backup = f"{backup_path}/{domain}/{datestring}/pvcdisks"
+    if not os.path.isdir(vm_target_backup):
+        try:
+            os.makedirs(vm_target_backup)
+        except Exception as e:
+            return False, f"ERROR: Failed to create backup directory: {e}"
+
+    # 5. Take snapshot of each disks with the name @backup_{datestring}
+    is_snapshot_create_failed = False
+    which_snapshot_create_failed = list()
+    msg_snapshot_create_failed = list()
+    for pool, volume, _ in vm_volumes:
+        retcode, retmsg = ceph.add_snapshot(zkhandler, pool, volume, snapshot_name)
+        if not retcode:
+            is_snapshot_create_failed = True
+            which_snapshot_create_failed.append(f"{pool}/{volume}")
+            msg_snapshot_create_failed.append(retmsg)
+
+    if is_snapshot_create_failed:
+        for pool, volume, _ in vm_volumes:
+            if ceph.verifySnapshot(zkhandler, pool, volume, snapshot_name):
+                ceph.remove_snapshot(zkhandler, pool, volume, snapshot_name)
+        return (
+            False,
+            f'ERROR: Failed to create snapshot for volume(s) {", ".join(which_snapshot_create_failed)}: {", ".join(msg_snapshot_create_failed)}',
+        )
+
+    # 6. Dump snapshot to folder with `rbd export` (full) or `rbd export-diff` (incremental)
+    is_snapshot_export_failed = False
+    which_snapshot_export_failed = list()
+    msg_snapshot_export_failed = list()
+    for pool, volume, _ in vm_volumes:
+        if incremental_parent is not None:
+            incremental_parent_snapshot_name = f"backup_{incremental_parent}"
+            retcode, stdout, stderr = common.run_os_command(
+                f"rbd export-diff --from-snap {incremental_parent_snapshot_name} {pool}/{volume}@{snapshot_name} {vm_target_backup}/{pool}.{volume}.{export_fileext}"
+            )
+            if retcode:
+                is_snapshot_export_failed = True
+                which_snapshot_export_failed.append(f"{pool}/{volume}")
+                msg_snapshot_export_failed.append(stderr)
+        else:
+            retcode, stdout, stderr = common.run_os_command(
+                f"rbd export --export-format 2 {pool}/{volume}@{snapshot_name} {vm_target_backup}/{pool}.{volume}.{export_fileext}"
+            )
+            if retcode:
+                is_snapshot_export_failed = True
+                which_snapshot_export_failed.append(f"{pool}/{volume}")
+                msg_snapshot_export_failed.append(stderr)
+
+    if is_snapshot_export_failed:
+        for pool, volume, _ in vm_volumes:
+            if ceph.verifySnapshot(zkhandler, pool, volume, snapshot_name):
+                ceph.remove_snapshot(zkhandler, pool, volume, snapshot_name)
+        return (
+            False,
+            f'ERROR: Failed to export snapshot for volume(s) {", ".join(which_snapshot_export_failed)}: {", ".join(msg_snapshot_export_failed)}',
+        )
+
+    # 7. Create and dump VM backup information
+    backup_type = "incremental" if incremental_parent is not None else "full"
+    vm_backup = {
+        "type": backup_type,
+        "datestring": datestring,
+        "incremental_parent": incremental_parent,
+        "retained_snapshot": retain_snapshot,
+        "vm_detail": vm_detail,
+        "backup_files": [
+            (f"pvcdisks/{p}.{v}.{export_fileext}", s) for p, v, s in vm_volumes
+        ],
+    }
+    with open(f"{vm_target_root}/{datestring}/pvcbackup.json", "w") as fh:
+        jdump(vm_backup, fh)
+
+    # 8. Remove snapshots if retain_snapshot is False
+    is_snapshot_remove_failed = False
+    which_snapshot_remove_failed = list()
+    msg_snapshot_remove_failed = list()
+    if not retain_snapshot:
+        for pool, volume, _ in vm_volumes:
+            if ceph.verifySnapshot(zkhandler, pool, volume, snapshot_name):
+                retcode, retmsg = ceph.remove_snapshot(
+                    zkhandler, pool, volume, snapshot_name
+                )
+                if not retcode:
+                    is_snapshot_remove_failed = True
+                    which_snapshot_remove_failed.append(f"{pool}/{volume}")
+                    msg_snapshot_remove_failed.append(retmsg)
+
+    tend = time.time()
+    ttot = round(tend - tstart, 2)
+    retlines = list()
+
+    if is_snapshot_remove_failed:
+        retlines.append(
+            f"WARNING: Failed to remove snapshot(s) as requested for volume(s) {', '.join(which_snapshot_remove_failed)}: {', '.join(msg_snapshot_remove_failed)}"
+        )
+
+    myhostname = gethostname().split(".")[0]
+    if retain_snapshot:
+        retlines.append(
+            f"Successfully backed up VM '{domain}' ({backup_type}@{datestring}, snapshots retained) to '{myhostname}:{backup_path}' in {ttot}s."
+        )
+    else:
+        retlines.append(
+            f"Successfully backed up VM '{domain}' ({backup_type}@{datestring}) to '{myhostname}:{backup_path}' in {ttot}s."
+        )
+
+    return True, "\n".join(retlines)
+
+
+def remove_backup(zkhandler, domain, backup_path, datestring):
+    tstart = time.time()
+
+    # 0. Validation
+    # Validate that VM exists in cluster
+    dom_uuid = getDomainUUID(zkhandler, domain)
+    if not dom_uuid:
+        return False, 'ERROR: Could not find VM "{}" in the cluster!'.format(domain)
+
+    # Validate that the source path is valid
+    if not re.match(r"^/", backup_path):
+        return (
+            False,
+            f"ERROR: Source path {backup_path} is not a valid absolute path on the primary coordinator!",
+        )
+
+    # Ensure that backup_path (on this node) exists
+    if not os.path.isdir(backup_path):
+        return False, f"ERROR: Source path {backup_path} does not exist!"
+
+    # Ensure that domain path (on this node) exists
+    vm_backup_path = f"{backup_path}/{domain}"
+    if not os.path.isdir(vm_backup_path):
+        return False, f"ERROR: Source VM path {vm_backup_path} does not exist!"
+
+    # Ensure that the archives are present
+    backup_source_pvcbackup_file = f"{vm_backup_path}/{datestring}/pvcbackup.json"
+    if not os.path.isfile(backup_source_pvcbackup_file):
+        return False, "ERROR: The specified source backup files do not exist!"
+
+    backup_source_pvcdisks_path = f"{vm_backup_path}/{datestring}/pvcdisks"
+    if not os.path.isdir(backup_source_pvcdisks_path):
+        return False, "ERROR: The specified source backup files do not exist!"
+
+    # 1. Read the backup file and get VM details
+    try:
+        with open(backup_source_pvcbackup_file) as fh:
+            backup_source_details = jload(fh)
+    except Exception as e:
+        return False, f"ERROR: Failed to read source backup details: {e}"
+
+    # 2. Remove snapshots
+    is_snapshot_remove_failed = False
+    which_snapshot_remove_failed = list()
+    msg_snapshot_remove_failed = list()
+    if backup_source_details["retained_snapshot"]:
+        for volume_file, _ in backup_source_details.get("backup_files"):
+            pool, volume, _ = volume_file.split("/")[-1].split(".")
+            snapshot = f"backup_{datestring}"
+            retcode, retmsg = ceph.remove_snapshot(zkhandler, pool, volume, snapshot)
+            if not retcode:
+                is_snapshot_remove_failed = True
+                which_snapshot_remove_failed.append(f"{pool}/{volume}")
+                msg_snapshot_remove_failed.append(retmsg)
+
+    # 3. Remove files
+    is_files_remove_failed = False
+    msg_files_remove_failed = None
+    try:
+        rmtree(f"{vm_backup_path}/{datestring}")
+    except Exception as e:
+        is_files_remove_failed = True
+        msg_files_remove_failed = e
+
+    tend = time.time()
+    ttot = round(tend - tstart, 2)
+    retlines = list()
+
+    if is_snapshot_remove_failed:
+        retlines.append(
+            f"WARNING: Failed to remove snapshot(s) as requested for volume(s) {', '.join(which_snapshot_remove_failed)}: {', '.join(msg_snapshot_remove_failed)}"
+        )
+
+    if is_files_remove_failed:
+        retlines.append(
+            f"WARNING: Failed to remove backup file(s) from {backup_path}: {msg_files_remove_failed}"
+        )
+
+    myhostname = gethostname().split(".")[0]
+    retlines.append(
+        f"Removed VM backup {datestring} for '{domain}' from '{myhostname}:{backup_path}' in {ttot}s."
+    )
+
+    return True, "\n".join(retlines)
+
+
+def restore_vm(zkhandler, domain, backup_path, datestring, retain_snapshot=False):
+    tstart = time.time()
+
+    # 0. Validations
+    # Validate that VM does not exist in cluster
+    dom_uuid = getDomainUUID(zkhandler, domain)
+    if dom_uuid:
+        return (
+            False,
+            f'ERROR: VM "{domain}" already exists in the cluster! Remove or rename it before restoring a backup.',
+        )
+
+    # Validate that the source path is valid
+    if not re.match(r"^/", backup_path):
+        return (
+            False,
+            f"ERROR: Source path {backup_path} is not a valid absolute path on the primary coordinator!",
+        )
+
+    # Ensure that backup_path (on this node) exists
+    if not os.path.isdir(backup_path):
+        return False, f"ERROR: Source path {backup_path} does not exist!"
+
+    # Ensure that domain path (on this node) exists
+    vm_backup_path = f"{backup_path}/{domain}"
+    if not os.path.isdir(vm_backup_path):
+        return False, f"ERROR: Source VM path {vm_backup_path} does not exist!"
+
+    # Ensure that the archives are present
+    backup_source_pvcbackup_file = f"{vm_backup_path}/{datestring}/pvcbackup.json"
+    if not os.path.isfile(backup_source_pvcbackup_file):
+        return False, "ERROR: The specified source backup files do not exist!"
+
+    # 1. Read the backup file and get VM details
+    try:
+        with open(backup_source_pvcbackup_file) as fh:
+            backup_source_details = jload(fh)
+    except Exception as e:
+        return False, f"ERROR: Failed to read source backup details: {e}"
+
+    # Handle incrementals
+    incremental_parent = backup_source_details.get("incremental_parent", None)
+    if incremental_parent is not None:
+        backup_source_parent_pvcbackup_file = (
+            f"{vm_backup_path}/{incremental_parent}/pvcbackup.json"
+        )
+        if not os.path.isfile(backup_source_parent_pvcbackup_file):
+            return (
+                False,
+                "ERROR: The specified backup is incremental but the required incremental parent source backup files do not exist!",
+            )
+
+        try:
+            with open(backup_source_parent_pvcbackup_file) as fh:
+                backup_source_parent_details = jload(fh)
+        except Exception as e:
+            return (
+                False,
+                f"ERROR: Failed to read source incremental parent backup details: {e}",
+            )
+
+    # 2. Import VM config and metadata in provision state
+    try:
+        retcode, retmsg = define_vm(
+            zkhandler,
+            backup_source_details["vm_detail"]["xml"],
+            backup_source_details["vm_detail"]["node"],
+            backup_source_details["vm_detail"]["node_limit"],
+            backup_source_details["vm_detail"]["node_selector"],
+            backup_source_details["vm_detail"]["node_autostart"],
+            backup_source_details["vm_detail"]["migration_method"],
+            backup_source_details["vm_detail"]["profile"],
+            backup_source_details["vm_detail"]["tags"],
+            "restore",
+        )
+        if not retcode:
+            return False, f"ERROR: Failed to define restored VM: {retmsg}"
+    except Exception as e:
+        return False, f"ERROR: Failed to parse VM backup details: {e}"
+
+    # 4. Import volumes
+    is_snapshot_remove_failed = False
+    which_snapshot_remove_failed = list()
+    msg_snapshot_remove_failed = list()
+    if incremental_parent is not None:
+        for volume_file, volume_size in backup_source_details.get("backup_files"):
+            pool, volume, _ = volume_file.split("/")[-1].split(".")
+            try:
+                parent_volume_file = [
+                    f[0]
+                    for f in backup_source_parent_details.get("backup_files")
+                    if f[0].split("/")[-1].replace(".rbdimg", "")
+                    == volume_file.split("/")[-1].replace(".rbddiff", "")
+                ][0]
+            except Exception as e:
+                return (
+                    False,
+                    f"ERROR: Failed to find parent volume for volume {pool}/{volume}; backup may be corrupt or invalid: {e}",
+                )
+
+            # First we create the expected volumes then clean them up
+            #   This process is a bit of a hack because rbd import does not expect an existing volume,
+            #   but we need the information in PVC.
+            #   Thus create the RBD volume using ceph.add_volume based on the backup size, and then
+            #   manually remove the RBD volume (leaving the PVC metainfo)
+            retcode, retmsg = ceph.add_volume(zkhandler, pool, volume, volume_size)
+            if not retcode:
+                return False, f"ERROR: Failed to create restored volume: {retmsg}"
+
+            retcode, stdout, stderr = common.run_os_command(
+                f"rbd remove {pool}/{volume}"
+            )
+            if retcode:
+                return (
+                    False,
+                    f"ERROR: Failed to remove temporary RBD volume '{pool}/{volume}': {stderr}",
+                )
+
+            # Next we import the parent images
+            retcode, stdout, stderr = common.run_os_command(
+                f"rbd import --export-format 2 --dest-pool {pool} {backup_path}/{domain}/{incremental_parent}/{parent_volume_file} {volume}"
+            )
+            if retcode:
+                return (
+                    False,
+                    f"ERROR: Failed to import parent backup image {parent_volume_file}: {stderr}",
+                )
+
+            # Then we import the incremental diffs
+            retcode, stdout, stderr = common.run_os_command(
+                f"rbd import-diff {backup_path}/{domain}/{datestring}/{volume_file} {pool}/{volume}"
+            )
+            if retcode:
+                return (
+                    False,
+                    f"ERROR: Failed to import incremental backup image {volume_file}: {stderr}",
+                )
+
+            # Finally we remove the parent and child snapshots (no longer required required)
+            if retain_snapshot:
+                retcode, retmsg = ceph.add_snapshot(
+                    zkhandler,
+                    pool,
+                    volume,
+                    f"backup_{incremental_parent}",
+                    zk_only=True,
+                )
+                if not retcode:
+                    return (
+                        False,
+                        f"ERROR: Failed to add imported image snapshot for {parent_volume_file}: {retmsg}",
+                    )
+            else:
+                retcode, stdout, stderr = common.run_os_command(
+                    f"rbd snap rm {pool}/{volume}@backup_{incremental_parent}"
+                )
+                if retcode:
+                    is_snapshot_remove_failed = True
+                    which_snapshot_remove_failed.append(f"{pool}/{volume}")
+                    msg_snapshot_remove_failed.append(retmsg)
+            retcode, stdout, stderr = common.run_os_command(
+                f"rbd snap rm {pool}/{volume}@backup_{datestring}"
+            )
+            if retcode:
+                is_snapshot_remove_failed = True
+                which_snapshot_remove_failed.append(f"{pool}/{volume}")
+                msg_snapshot_remove_failed.append(retmsg)
+
+    else:
+        for volume_file, volume_size in backup_source_details.get("backup_files"):
+            pool, volume, _ = volume_file.split("/")[-1].split(".")
+
+            # First we create the expected volumes then clean them up
+            #   This process is a bit of a hack because rbd import does not expect an existing volume,
+            #   but we need the information in PVC.
+            #   Thus create the RBD volume using ceph.add_volume based on the backup size, and then
+            #   manually remove the RBD volume (leaving the PVC metainfo)
+            retcode, retmsg = ceph.add_volume(zkhandler, pool, volume, volume_size)
+            if not retcode:
+                return False, f"ERROR: Failed to create restored volume: {retmsg}"
+
+            retcode, stdout, stderr = common.run_os_command(
+                f"rbd remove {pool}/{volume}"
+            )
+            if retcode:
+                return (
+                    False,
+                    f"ERROR: Failed to remove temporary RBD volume '{pool}/{volume}': {stderr}",
+                )
+
+            # Then we perform the actual import
+            retcode, stdout, stderr = common.run_os_command(
+                f"rbd import --export-format 2 --dest-pool {pool} {backup_path}/{domain}/{datestring}/{volume_file} {volume}"
+            )
+            if retcode:
+                return (
+                    False,
+                    f"ERROR: Failed to import backup image {volume_file}: {stderr}",
+                )
+
+            # Finally we remove the source snapshot (not required)
+            if retain_snapshot:
+                retcode, retmsg = ceph.add_snapshot(
+                    zkhandler,
+                    pool,
+                    volume,
+                    f"backup_{datestring}",
+                    zk_only=True,
+                )
+                if not retcode:
+                    return (
+                        False,
+                        f"ERROR: Failed to add imported image snapshot for {volume_file}: {retmsg}",
+                    )
+            else:
+                retcode, stdout, stderr = common.run_os_command(
+                    f"rbd snap rm {pool}/{volume}@backup_{datestring}"
+                )
+                if retcode:
+                    return (
+                        False,
+                        f"ERROR: Failed to remove imported image snapshot for {volume_file}: {stderr}",
+                    )
+
+    # 5. Start VM
+    retcode, retmsg = start_vm(zkhandler, domain)
+    if not retcode:
+        return False, f"ERROR: Failed to start restored VM {domain}: {retmsg}"
+
+    tend = time.time()
+    ttot = round(tend - tstart, 2)
+    retlines = list()
+
+    if is_snapshot_remove_failed:
+        retlines.append(
+            f"WARNING: Failed to remove hanging snapshot(s) as requested for volume(s) {', '.join(which_snapshot_remove_failed)}: {', '.join(msg_snapshot_remove_failed)}"
+        )
+
+    myhostname = gethostname().split(".")[0]
+    retlines.append(
+        f"Successfully restored VM backup {datestring} for '{domain}' from '{myhostname}:{backup_path}' in {ttot}s."
+    )
+
+    return True, "\n".join(retlines)

debian/changelog

@@ -1,3 +1,25 @@
+pvc (0.9.80-0) unstable; urgency=high
+
+  * [CLI] Improves CLI performance by not loading "pkg_resources" until needed
+  * [CLI] Improves the output of the audit log (full command paths)
+  * [Node Daemon/API Daemon] Moves the sample YAML configurations to /usr/share/pvc instead of /etc/pvc and cleans up the old locations automatically
+  * [CLI] Adds VM autobackup functionality to automate VM backup/retention and scheduling
+  * [CLI] Handles the internal store in a better way to ensure CLI can be used as a module properly
+
+ -- Joshua M. Boniface <joshua@boniface.me>  Fri, 27 Oct 2023 09:56:31 -0400
+
+pvc (0.9.79-0) unstable; urgency=high
+
+  **API Changes**: New endpoints /vm/{vm}/backup, /vm/{vm}/restore
+
+  * [CLI Client] Fixes some storage pool help text messages
+  * [Node Daemon] Increases the IPMI monitoring plugin timeout
+  * [All] Adds support for VM backups, including creation, removal, and restore
+  * [Repository] Fixes shebangs in scripts to be consistent
+  * [Daemon Library] Improves the handling of VM list arguments (default None)
+
+ -- Joshua M. Boniface <joshua@boniface.me>  Tue, 24 Oct 2023 02:10:24 -0400
+
 pvc (0.9.78-0) unstable; urgency=high
 
   * [API, Client CLI] Fixes several bugs around image uploads; adds a new query parameter for non-raw images

debian/pvc-client-cli.install

@@ -0,0 +1 @@
+client-cli/autobackup.sample.yaml usr/share/pvc

debian/pvc-daemon-api.install

@@ -1,7 +1,7 @@
 api-daemon/pvcapid.py usr/share/pvc
 api-daemon/pvcapid-manage*.py usr/share/pvc
 api-daemon/pvc-api-db-upgrade usr/share/pvc
-api-daemon/pvcapid.sample.yaml etc/pvc
+api-daemon/pvcapid.sample.yaml usr/share/pvc
 api-daemon/pvcapid usr/share/pvc
 api-daemon/pvcapid.service lib/systemd/system
 api-daemon/pvcapid-worker.service lib/systemd/system

debian/pvc-daemon-api.postinst

@@ -18,3 +18,6 @@ fi
 if [ ! -f /etc/pvc/pvcapid.yaml ]; then
     echo "NOTE: The PVC client API daemon (pvcapid.service) and the PVC provisioner worker daemon (pvcapid-worker.service) have not been started; create a config file at /etc/pvc/pvcapid.yaml, then run the database configuration (/usr/share/pvc/pvc-api-db-upgrade) and start them manually."
 fi
+
+# Clean up any old sample configs
+rm /etc/pvc/pvcapid.sample.yaml || true

debian/pvc-daemon-node.install

@@ -1,5 +1,5 @@
 node-daemon/pvcnoded.py usr/share/pvc
-node-daemon/pvcnoded.sample.yaml etc/pvc
+node-daemon/pvcnoded.sample.yaml usr/share/pvc
 node-daemon/pvcnoded usr/share/pvc
 node-daemon/pvcnoded.service lib/systemd/system
 node-daemon/pvc.target lib/systemd/system

debian/pvc-daemon-node.postinst

@@ -14,3 +14,6 @@ if systemctl is-active --quiet pvcnoded.service; then
 else
     echo "NOTE: The PVC node daemon (pvcnoded.service) has not been started; create a config file at /etc/pvc/pvcnoded.yaml then start it."
 fi
+
+# Clean up any old sample configs
+rm /etc/pvc/pvcnoded.sample.yaml || true

gen-api-migrations

@@ -1,4 +1,4 @@
-#!/bin/bash
+#!/usr/bin/env bash
 
 # Generate the database migration files
 

gen-zk-migrations

@@ -1,4 +1,4 @@
-#!/bin/bash
+#!/usr/bin/env bash
 
 # Generate the Zookeeper migration files
 

node-daemon/plugins/ipmi

@@ -76,7 +76,7 @@ class MonitoringPluginScript(MonitoringPlugin):
         ipmi_password = self.config["ipmi_password"]
         retcode, _, _ = run_os_command(
             f"/usr/bin/ipmitool -I lanplus -H {ipmi_hostname} -U {ipmi_username} -P {ipmi_password} chassis power status",
-            timeout=2
+            timeout=5
         )
 
         if retcode > 0:

node-daemon/plugins/nics

@@ -132,7 +132,7 @@ class MonitoringPluginScript(MonitoringPlugin):
                 for slave_interface in slave_interfaces:
                     if slave_interface[1] == 'up':
                         slave_interface_up_count += 1
-                if slave_interface_up_count < 2:
+                if slave_interface_up_count < len(slave_interfaces):
                     messages.append(f"{dev} DEGRADED with {slave_interface_up_count} active slaves")
                     health_delta += 10
                 else:

node-daemon/pvcnoded/Daemon.py

@@ -49,7 +49,7 @@ import re
 import json
 
 # Daemon version
-version = "0.9.78"
+version = "0.9.80"
 
 
 ##########################################################

node-daemon/pvcnoded/util/services.py

@@ -77,5 +77,5 @@ def start_system_services(logger, config):
     start_ceph_mon(logger, config)
     start_ceph_mgr(logger, config)
 
-    logger.out("Waiting 3 seconds for daemons to start", state="s")
-    sleep(3)
+    logger.out("Waiting 10 seconds for daemons to start", state="s")
+    sleep(10)